WO2021088539A1 - Tilt-shift stereo camera - Google Patents

Abstract

A tilt-shift stereo camera. The left and right lens groups in a tilt-shift stereo camera that are independent of each other, are identical, and have centerlines arranged parallel to each other, are translated close to each other in the horizontal direction by a distance of L=T÷(2A). No trapezoidal distortion occurs in two images with different angles of view obtained by the tilt-shift stereo camera, and the relationship between the stereo depth of an object of interest in a real scene and the stereo depth of the convergent point of a stereo image of the object of interest is a linear relationship. Also disclosed is a screen-mapping chip, which solves the problem of separation between the focal plane of eyes and the image plane of the stereo image, and a healthy stereo player, which makes the stereo image measurement process smarter, simpler, and more accurate. The present application can be used in medical endoscopes, laparoscopes, industrial endoscopes, stereo cameras, stereo measurement and other stereo image production and application fields.

Description

Axis-shifting stereo camera Technical field

The invention relates to a dual-lens stereo camera, a stereo image linear optical design, an eye focal plane and a stereo image image plane coincidence technology, a stereo image measurement technology and a stereo image positioning and tracking technology.

Background technique

Two cameras usually use the convergence method or the parallel method to shoot stereo images. The three-dimensional image effect obtained by the convergence method is the same as the way and effect of the eye observing an object of interest, but the two images obtained have the problem of trapezoidal distortion. The way and effect of an object of interest obtained by the parallel method is the same as the way and effect of an eye observing an object of interest located at infinity. However, when the distance of the object of interest is limited, the three-dimensional effect of the three-dimensional image of the object of interest obtained by the parallel method is different from the effect of observing the object of interest by eyes, and it is not an ideal three-dimensional representation.

The image obtained by the equivalent convergence method is not only consistent with the way the eyes observe the world, but the relationship between the stereo depth of an object of interest and the stereo depth of the convergent point of the stereo image of the object of interest in the real scene can be expressed by a set of linear mathematical formulas. However, the complete optical imaging of a core-shifting stereo camera with an image sensor poses a challenge, because an image sensor cannot divide and translate.

Modern image sensor technology continues to introduce image sensors with ultra-high-definition image resolution, which exceeds the requirements of stereoscopic image transmission and playback standards. A stereo camera with an image sensor can not only meet the resolution, transmission and playback standards of stereo images, but also solve the synchronization problem of stereo images. However, the stereoscopic image obtained on an image sensor needs to be translated by an algorithm to have linear characteristics.

At present, all stereo players are a display technology based on flat screens. The biggest problem with this technology is that the focal plane of the eye is separated from the image plane of the stereoscopic image. This is one of the main reasons why human eyes feel fatigue and physical discomfort after watching the stereoscopic image for a period of time.

The mainstream binocular stereoscopic image measurement method is a technology based on the convergence method, but the trapezoidal distortion and trigonometric function existing in the convergence method cause additional errors to the stereoscopic image measurement results.

The axis-shifting stereo camera proposed by the present invention solves the above-mentioned problems in different application fields, and has the characteristics of simple operation, low cost, easy promotion and popularization.

Summary of the invention

The purpose of the present invention is to provide an axis-shifting stereo camera. First, it solves the technical problem that the relationship between the stereo depth of an object of interest and the stereo depth of the convergent point of the stereo image of the object of interest in the real scene satisfies a set of linear mathematical formulas; Second, it solves the technical problem of separating the focal plane of the eye from the image plane of the stereo image; third, it solves the technical problem of turning a stereo player into a healthy stereo player; fourth, it solves the problem of stereo image measurement The process is smarter, simpler, and the result is more accurate technical issues.

An axis-shifting stereo camera is composed of two independent, identical and centrally parallel lens groups and one or two identical image sensors CCD or CMOS. The two lens groups can move along a straight line which is located on a plane formed by the center lines of the two lens groups and is perpendicular to the center lines of the two lens groups, respectively, and translate L=T÷(2A) in opposite directions. distance. When shifting the axis, the position of one or two image sensors does not change. Among them, T is the distance between human eyes and A is the magnification of the screen.

The two lens groups in a shift stereo camera are exactly the same, including (not limited to) the focal length, angle of view, aperture, center position of the lens group, lens design and type in the lens group, lens number, lens material, and lens surface coating , The inner surface coating, optical design, structural design and other parameters of the lens module housing.

The distance between two lens groups in an axis-shifting stereo camera changes due to the axis-shifting, but the position of one or two image sensors in the axis-shifting stereo camera remains unchanged. An axis-shifting stereo camera can be divided into two designs: independent viewing distance and non-independent viewing distance. The independent viewing distance design is the changing process of the viewing distance t of an axis-shifting stereo camera and the axis-shifting process is two independent processes and operations. A typical example is an optical steering system, such as a pair of steering right-angle prisms, respectively provided in the two lens groups of an axis-shifting stereo camera. A pair of steering right-angle prisms divide a lens group into two parts: the front objective lens group and the rear imaging lens group. The center lines of a pair of steering right-angle prisms are located on a plane formed by the center lines of the two lens groups in an axis-shifting stereo camera. The center line of the right-angle exit surface of the front right-angle prism and the center line of the right-angle entrance surface of the rear right-angle prism and The center lines of the two lens groups are vertical. In a lens group, the image collected by the objective lens group enters the front right-angle prism, the inclined surface of the front right-angle prism bends the image by 90° and then enters the rear right-angle prism, and the rear right-angle prism bends the image again by 90° and then enters To the imaging lens group, the imaging lens group projects the image onto the imaging surface of the image sensor for imaging. When the axis is shifted, the rear right-angle prism and the imaging lens group in the two lens groups simultaneously or respectively move toward each other by a distance of L=T÷(2A). The distance between the two objective lens groups in the two lens groups and the front right-angle prism remains unchanged. The design of non-independent viewing distance is that the distance between the center lines of the two lens groups of a pivot-shifting stereo camera is equal to the viewing distance t of the pivot-shifting stereo camera regardless of whether it is before or after pivoting. The three-dimensional effect of the three-dimensional image obtained by an axis-shifting stereo camera with an independent visual distance design has more expression methods and quality.

After the axis is shifted, the viewing distance t of an axis-shifting stereo camera is between 3 mm and 200 mm.

An axis-shift stereo camera outputs two different axis-shift image formats, one axis-shift left-right format and two independent axis-shift image formats. For an axis-shifting stereo camera provided with an image sensor, the center lines of the left and right lens groups respectively pass through the center of the left half and the center of the right half of the imaging surface of an image sensor. When the axis is shifted, the two lens groups are respectively translated along the horizontal direction toward the opposite direction by a distance of L=T÷(2A)=(T×w)÷(4W). After the axis is shifted, the left and right images collected by the left and right lens groups are respectively imaged on the left half and the right half of the imaging surface of an image sensor and output an image in a shifted left and right format. An image in the shift left and right format is composed of a left image and a right image arranged in a left-to-right arrangement to form a complete image.

For an axis-shifting stereo camera provided with two image sensors, the center lines of the left and right lens groups respectively pass through the centers of the imaging surfaces of the left and right image sensors. When the axis is shifted, the left and right lens groups are respectively translated along the horizontal direction toward the opposite direction by a distance of L=T÷(2A)=(T×w)÷(2W). After the axis is shifted, the left and right images collected by the left and right lens groups are respectively imaged on the imaging surfaces of the left and right independent image sensors and output two independent left and right images. Among them, w is the horizontal length of the effective imaging surface of the image sensor, and W is the horizontal length of the effective playback surface of the screen.

A shift-left-right format image and two independent shift-axis images have advantages over traditional left-right format images and two independent images; first, in the shift-shift format image, one focus on the object in the real scene There is a linear relationship between the stereo depth and the stereo depth of the convergent point of the stereo image of the object of interest; second, an object of interest in the real scene corresponds to the only stereo image of the object of interest without distortion; third, one is located at the center of the stereo camera The object of interest on the axis is imaged at the center of the effective imaging surface of the image sensor.

After shifting, the smallest imaging circle when the two images obtained by the two lens groups meet the required image resolution format is the smallest shifting imaging circle of the two lens groups. The smallest axis-shifting imaging circle diameters of the two lens groups in an axis-shifting stereo camera are the same. For an axis-shifting stereo camera equipped with an image sensor, the smallest axis-shifting imaging circle diameter of the two lens groups is D _min =2√[(w/4+L) ² +(g/2) ² ]. For an axis-shifting stereo camera equipped with two image sensors, the minimum axis-shifting imaging circle diameter of the two lens groups is D _min =2√[(w/2+L) ² +(g/2) ² ]. Among them, g is the vertical height of the required image resolution format.

Two cameras usually use two shooting methods when shooting stereo images, the convergence method and the parallel method. The three-dimensional image effect obtained by using the convergence method to shoot an object of interest is the same as the way and effect of the eye observing an object of interest. When the left and right cameras converge the centerline of the lens group on an object of interest located on the central axis of the stereo camera, the left and right images collected by the left and right cameras are respectively imaged at the center of the effective imaging surfaces of the left and right image sensors. However, the left and right images obtained have trapezoidal distortion and cannot be perfectly blended. Using the parallel method to photograph an object of interest is the same as the way the eyes observe an object of interest located at infinity, and the obtained image has no keystone distortion. However, for an object of interest located at a limited distance, the three-dimensional image of the object of interest obtained by the parallel method is different from the way the eyes observe an object of interest, and the stereoscopic effect of the screen is not an ideal way of stereoscopic expression.

The principle of equivalent convergence is that in a stereo camera composed of two independent lens groups with the same centerline and parallel to each other, the two lens groups or two image sensors are along a plane formed by the center lines of the two lens groups. The upper part and the linear direction perpendicular to the center lines of the two lens groups are translated equally, so that two images of an object of interest on the central axis of the stereo camera collected by the two lens groups are respectively on the effective imaging surfaces of the two image sensors Center imaging. The equivalent convergence method is a stereoscopic shooting method based on the equivalent convergence principle. Before shooting, put two lens groups or two image sensors in a stereo camera composed of two independent lens groups with the same and center lines parallel to each other along a plane formed by the center lines of the two lens groups. And it is translated by a distance of L=T÷(2A) or h=T÷(2A) in a straight line direction perpendicular to the center lines of the two lens groups. The three-dimensional effect of a three-dimensional image of an object of interest obtained by using the equivalent convergence method is the same as the three-dimensional effect of a three-dimensional image of an object of interest obtained by the convergence method, but there is no trapezoidal distortion in the two images. In fact, the most important significance of the principle of equivalent convergence and the use of the equivalent convergence method is to establish a linear relationship between the stereo depth of an object of interest in the real scene and the stereo depth of the convergent point of the stereo image of the object of interest. The physical meaning is that the three-dimensional image of a focus, a focus line and a focus plane corresponding to a focus point, a focus line and a focus plane in the real scene is unique and undistorted.

According to the above-mentioned [0016], using the equivalent convergence method to shoot an object of interest in a real scene is a sufficient and necessary condition for a linear relationship between the stereo depth of the object of interest in the real scene and the stereo depth of the stereo image convergence point of the object of interest. Two kinds of stereo cameras are designed based on the principle of equivalent convergence; the first is an axis-shifting stereo camera. Before shooting, the two lens groups of an axis-shifting stereo camera are located on a plane formed by the center lines of the two lens groups and are perpendicular to the center lines of the two lens groups, facing each other in the opposite direction. Translation L=T÷(2A) distance. When panning, the position of one or two image sensors in the axis-shifting stereo camera remains unchanged. After the axis is shifted, an axis-shifting stereo camera images an object of interest located on the central axis at the center of the effective imaging surfaces of the two image sensors. For a stereo camera equipped with an image sensor, an axis-shift stereo camera is an ideal optical design and solution. The second type is a core-moving stereo camera. Before shooting, the two image sensors in a core-shifting stereo camera are located on a plane formed by the center lines of the two lens groups and are perpendicular to the center lines of the two lens groups, and face in opposite directions. Translation h=T÷(2A) distance. When panning, the positions of the two lens groups in the core-shifting stereo camera remain unchanged. After the core is moved, a core-moving stereo camera images an object of interest located on the central axis at the center of the effective imaging surfaces of the two image sensors. In the above two types of stereo cameras, the translation formulas L=T÷(2A) and h=T÷(2A), the coordinate system and the origin of the coordinate system are all the same. However, the meaning of t in the two different stereo cameras is different. The t in the tilt-shift stereo camera is the distance between the center lines of the two lens groups after the tilt. The t in the core-shifting stereo camera is the distance between the center lines of the left and right lens groups. The relationship between shifting axis and shifting core is L=t×h÷(t+2h).

A kind of stereo image translation command is based on the principle of equivalent convergence, the two images collected by a stereo camera composed of two independent lens groups or cameras arranged in parallel with the center line are located along a line in the two lens groups or On a plane formed by the center line of the camera and in a straight line direction perpendicular to the two lens groups or the center line of the camera, the distance h=T÷(2A) is respectively translated toward the opposite direction. After translation, the three-dimensional effect of the two images is the same as the three-dimensional effect of the two images obtained by the equivalent convergence method. For a core-moving stereo camera equipped with an image sensor, because an image sensor cannot be divided and translated, a stereoscopic image translation instruction provides an optical alternative solution for a core-moving stereo camera with an image sensor. A three-dimensional image translation instruction can also be applied to a pivot-shifting stereo camera and a core-shifting stereo camera with two image sensors. There are many ways of image translation. The following example is just one of them and explains the image translation in principle. For a left-right format image; in the first step, a vertical line intersecting the right edge of the left image and the left edge of the right image in the left-right format image is used as a dividing line. In the left image, the left image is cut along a vertical straight line with a distance dividing line h=T÷(2A)=(T×w)÷(4W), and the left image of the cut vertical straight line is retained. In the right image, the right image is cut along a vertical straight line with a distance dividing line h=T÷(2A)=(T×w)÷(4W), and the image on the right of the cut vertical straight line is retained. The second step is to move the retained left image to the right by the distance h=T÷(2A)=(T×w)÷(4W). Move the retained right image to the left by a distance of h=T÷(2A)=(T×w)÷(4W). The left and right images are re-spliced into a new left and right format image. In a new left-right format image, the left edge of the left image and the right edge of the right image have two vertical image blank areas each with a width of h. For two independent images on the left and right; in the first step, in the left image, cut the left image along a vertical straight line from the right edge of h=T÷(2A)=(T×w)÷(2W) , Keep the image on the left of the vertical straight line after cutting. In the right image, the right image is cut along a vertical straight line that is h=T÷(2A)=(T×w)÷(2W) from the left edge, and the image on the right of the cut vertical straight line is retained. This translation method causes two vertical image blank areas with a width of h at the left edge of the left image and the right edge of the right image respectively. Compared with the axis-shifting stereo camera and the core-shifting stereo camera described in [0017], the advantages of a stereo image translation instruction are (not limited to); first, it solves the core-shifting stereo camera with an image sensor The problem that the image sensor cannot be translated; secondly, after translation, the stereoscopic effect of the stereoscopic image is the same as the stereoscopic effect of the stereoscopic image obtained by the shifting or core-shifting stereo camera; thirdly, it can not only be applied to the shifting and core-shifting stereo cameras, but also It can also be applied to all stereo cameras composed of two independent left and right lens groups or cameras arranged in parallel with the center line; fourth, it can be applied not only to stereo cameras equipped with an image sensor, but also to stereo cameras equipped with an image sensor. A stereo camera with two image sensors; Fifth, for the shooting needs of frequently changing objects of interest, the process of resetting a new object of interest is simple, easy to operate and easy to use; Sixth, the stereo image convergence of different objects of interest can be changed at any time Click to obtain the three-dimensional effect and expression mode that changes the original scene of the entire three-dimensional image. However, the shortcomings of this technology are also obvious; first, after translation, the image of a vertical area where the left and right outer edges of the image have a width of h is cut, which is equivalent to reducing the angle of view of the lens group; second, Cause image delay.

A shifting device is to move the two lens groups in a shifting stereo camera along a straight line which is located on a plane formed by the center lines of the two lens groups and is perpendicular to the center lines of the two lens groups, respectively facing each other. A device that translates in the direction of L=T÷(2A). For an axis-shifting stereo camera equipped with an image sensor, the translation of each lens group is L=(T×w)÷(4W) distance. For an axis-shifting stereo camera equipped with two image sensors, the translation of each lens group is L=(T×w)÷(2W) distance. When shifting the axis, the positions of all image sensors in the axis-shifting stereo camera remain unchanged. A shifting device has two different shifting setting modes. The first setting mode is fixed. After the terminal stereo player has been determined, an axis-shifting stereo camera can preset the axis-shifting amount L required by the two optical lens groups before packaging. This lens group has a fixed amount of shifting, so there is no need to set up a shifting device, but the stereo images obtained by this shifting stereo camera need to be played in a certain stereo player to get the best stereo effect . If the terminal stereo player changes, a kind of stereo image translation instruction can be used to perform additional compensation on the shift axis L to obtain an ideal stereo effect. The second setting mode is adjustable. An axis-shifting device is provided with an axis-shifting fine-tuning mechanism and knob with original zero point and scale. Adjust the knob on the fine-tuning mechanism to change the distance between the two lens groups synchronously. When the knob rotates in one direction, the two lens groups translate in the opposite direction to each other. When the knob is rotated in the opposite direction, the two lens groups are translated in the opposite direction to each other. Because the change in the distance between the two lens groups is very small, the lens shift device is a precise fine-tuning device.

An image processor is one equipped with one or two image processing chips ISP, one or two wireless modules, an image synchronizer, a touch screen, a data memory and an operating system, and also includes an integrated and stored multiple instructions , A device on the same screen chip loaded and executed by the processor.

The number of image processing chips set in an image processor is the same as the number of image sensors in an axis-shifting stereo camera. For an axis-shift stereo camera equipped with an image sensor, an image processing chip processes, corrects and optimizes the left and right images in the axis-shift left and right format from one image sensor. For an axis-shifting stereo camera equipped with two image sensors, two image processing chips respectively process, correct and optimize two independent axis-shift images from the two image sensors. This correction, processing and optimization include (not limited to); white balance, increase color saturation, increase sharpness, brightness, contrast, reduce noise, image edge and detail restoration, compression and other parameters.

One or two wireless communication modules in an image processor respectively output the images, pictures, voices and texts corrected, processed and optimized by the image processing chip processor to the stereo player, touch screen, and remote in real time. The control center, database, and other third parties can also interact and communicate with third parties in real-time multimedia.

A touch screen in an image processor provides a human-computer interaction interface of an operating system. The operation methods include touch screen pen, finger, mouse and keyboard. A touch screen can be a traditional touch screen or a stereo touch screen. An operating system realizes human-computer interaction through a touch screen and operating interface, operating instructions to manage pages and images, image input, output, storage, loading, and execution of instructions for the integration and storage of a chip on the same screen. After correction, processing, optimization and shifting, the left and right shifting format or two independent shifting images are output to the stereo player, touch screen, remote control center and database, open interface and other operating systems and third-party application software Compatible, download various applications and APP links, other third parties, and can interact and communicate with other third parties in real-time multimedia.

An on-screen chip in an image processor is an integrated and stored three-dimensional image translation instruction, a three-dimensional image measurement instruction, a three-dimensional image positioning tracking instruction, a three-dimensional image on-screen instruction and an equivalent convergence Click the reset instruction chip. A same-screen chip is set in the image processor as an application chip, and the processor loads and executes the functions of stereo image positioning, matching, tracking, measurement, equivalent convergence point reset and the same screen.

The origin (0', 0', 0') of a three-dimensional image acquisition space coordinate system (x', y', z') is located at the midpoint of the line connecting the two centerlines of the camera lens that are arranged parallel to each other. The origin (0”, 0”, 0”) of the coordinate system (x”, y”, z”) of a stereoscopic video playback space is located at the midpoint of the line connecting the human eyes. Place a stereoscopic image acquisition space coordinate system (x',y',z') and a stereoscopic image playback space coordinate system (x”,y”,z”) together, and place the origin of the two coordinate systems (0 ',0',0') and (0”,0”,0”) overlap to form a new coordinate system (x,y,z) and (0,0,0). In the new coordinate system, the relationship between the stereo depth of an object of interest in the real scene collected by an axis-shifting stereo camera and the stereo depth of the convergent point of the stereo image of the object of interest is Z _C =Z _D ×[T÷(A×F× t)]×Z. The formula shows that the relationship between the stereo depth Z of an object of interest in the real scene and the stereo depth Z _{C of the} convergent point of the stereo image of the object of interest is a linear relationship. In the formula, Z _D is the distance from the origin of the coordinate system to the flat screen, Z is the stereo depth of an object of interest in the real scene, and Z _C is the stereo depth of the convergence point of the stereo image of the object of interest.

All the mainstream stereoscopic image display technologies in the current market are based on the principle of convergence of stereoscopic images on flat screens. When the left and right images of an object of interest collected by the left and right cameras with different perspectives are projected onto a flat screen at the same time and the left and right eyes can only see the left and right images on the screen, the brain is aligned with the left and right images. The left and right images with different perspectives seen by the eyes are merged to feel a three-dimensional image.

In real life, when observing an object of interest, the eyes will automatically converge on the object of interest. After the brain merges two images with different perspectives obtained by the eyes, a three-dimensional image felt by the brain appears on the object of interest. In a flat screen playback system, a person's left and right eyes focus on the left and right images on the flat screen, respectively, so the flat screen is the focus plane of the eyes. Z _D is a constant. According to the experience in real life, the eyes are focused on the left and right images on a flat screen. After the brain fusion, the convergence point of the two images should also appear on the screen, Z _C =Z _{D. But the formula Z C} =Z _D ×[T÷(A×F×t)]×Z stated in the above [0025] _{indicates that Z C is} not equal to Z _D , or the image plane of the focal plane of the eye and the converging point of the stereoscopic image It does not overlap. This phenomenon is one of the fundamental reasons that cause the eyes to feel fatigue, dizziness and physical discomfort after watching stereo images for a period of time.

A three-dimensional image same-screen instruction is based on the principle of equivalent convergence. When the relationship between the screen magnification A and the three-dimensional depth Z of an object of interest in the real scene changes according to the formula A=[T÷(F×t)]×Z, A convergent point of a three-dimensional image of an object of interest in a real scene collected by a stereo camera is always kept on the screen. _{The formula Z C} =Z _D ×[T÷(A×F×t)]×Z stated in the above [0025] shows that the necessary and sufficient condition for the coincidence of the focal plane of the human eye and the image plane of the stereo image is [T÷ (A×F×t)]×Z=1 or A=[T÷(F×t)]×Z=k×Z, where k=T÷(F×t) is a constant. When the three-dimensional depth coordinate Z of an object of interest in the real scene changes by ΔZ, ΔA=k×ΔZ. According to the definition, A=W/w, ΔA=W/Δw, Δw=W÷(k×ΔZ). In the formula, the parameter W is a constant, and the parameter w is regarded as a variable. When the distance Z between an object of interest and the camera changes in the real scene, w will be equivalently changed synchronously. The equivalent result of this change is that the stereoscopic image on the playback screen is enlarged or reduced, which is equivalent to the zooming process of a zoomable lens. When an object of interest in the real scene becomes farther away from the camera, ΔZ>0, then ΔA>0, Δw<0, which is equivalent to that the focal length of the stereo camera becomes larger, the viewing angle becomes smaller, and the image on the image sensor becomes smaller, so the screen The images in became smaller and smaller. The visual effect looks like a three-dimensional image equivalent to an object of interest in the real scene becomes more and more distant on the screen. In the same way, when an object of interest in the real scene is closer to the camera, ΔZ<0, then ΔA<0, Δw>0, which is equivalent to the focal length of the stereo camera becomes smaller, the viewing angle becomes larger, and the image on the image sensor becomes larger, so The image on the screen becomes bigger and bigger. The visual effect looks like a three-dimensional image equivalent to an object of interest in the real scene becomes closer and closer on the screen. The changing method, process and perspective effect of the image on the screen are consistent with the way human eyes observe an object of interest in the real scene, experience and perspective effect. The above description is a qualitative description of the change of the image magnification A in order to meet the same screen condition. Specific and clear quantitative results of ΔA need to introduce the concept of parallax between the two images, and the detailed derivation will be derived in the following description. The vertical magnification of the screen is B=V/v, where V is the vertical height of the effective playback surface of the screen, and v is the vertical height of the effective imaging surface of the image sensor. When the three-dimensional depth of an object of interest in the real scene changes by ΔZ, the three-dimensional image on the screen will also be enlarged or reduced, and the magnification change rate in the horizontal and vertical directions of the screen is equal, ΔB=ΔA.

For an axis-shifting stereo camera, the screen magnification A can be used to determine or change the space coordinates (0,0,Z _conv ) of the equivalent convergence point M of an axis-shifting stereo camera, where Z _conv = (F×t )÷(2L)=(A×F×t)÷T=C×A, where C=(F×t)÷T=1/k is a constant. Because L=T÷(2A), the same result can be obtained by changing A or L. When the equivalent convergence point M of an axis-shifting stereo camera is set on an object of interest, the spatial coordinates of the object of interest are (0,0,Z=Z _conv ). When the left and right images of the object of interest are projected on the screen, the images of the left and right images on the flat screen are superimposed, and the three-dimensional images of the object of interest appear in the brain. At this time, the two left and right images of the object of interest appear on the screen. The parallax of the image is zero. When the equivalent convergence point M of a stereo camera is set behind an object of interest, the spatial coordinates of the object of interest are (0,0,Z>Z _conv ). When the left and right images of the object of interest are projected on the screen, the brain feels that the three-dimensional image of the object of interest appears behind the screen. At this time, the parallax of the left and right images is positive. When the equivalent convergence point M of a stereo camera is set between an object of interest and the stereo camera, the spatial coordinates of the object of interest are (0,0,Z<Z _conv ). When the left and right images of the object of interest are projected on the screen, the three-dimensional images of the object felt in the brain appear between the screen and the audience. At this time, the parallax of the left and right images of the object of interest is negative.

The translation formula L=T÷(2A) shows that L has nothing to do with t. For an axis-shifting stereo camera designed with independent viewing distance, the distance t between the rear right-angle prism in the two lens groups and the imaging lens group can be based on the formula Z _conv = (A×F× t)÷T is determined.

The stereo depth magnification ratio of an object of interest and the corresponding stereo image is η=(Z _c2 －Z _c1 )÷(Z ₂ －Z ₁ )=(Z _D ×T)÷(A×F×t)=(Z _D /Z _conv ). The formula shows that the stereo depth magnification ratio η is proportional to the distance between the eyes and the screen.

According to Gauss's law and the definition of the lateral magnification of the camera lens:

m=x′/x=y′/y=s′/s

Among them, s'=F×(1-m) is the image distance, and s=F×(1/m-1) is the object distance. The horizontal magnification of a three-dimensional image of an object of interest on the screen is m×A (x and y directions).

According to the definition of the longitudinal magnification of the camera lens:

与屏幕放大率A无关，因为公式中使用m×A代替m。让：

In the above formula, s1 and s2 are the depth coordinates _{of the front and rear end faces of an object of interest in the real scene respectively, and m 1} and m ₂ are the lateral magnifications of the lens at the front and rear end faces of an object of interest in the real scene. . In a linear space, according to the definition of image magnification, the lateral magnification has nothing to do with the position of the object of interest, or m = m ₁ = m ₂ . The above formula also shows; the longitudinal magnification of the camera lens

It has nothing to do with the screen magnification A, because m×A is used instead of m in the formula. yield:

Get Z _D ×[T÷(A×F×t)]=(Z _D /Z _conv )=m ²

The formula η=(Z _D /Z _conv )=m ² or Z _D =m ² ×Z _conv . The formula surface, when the distance between the human eye and the stereo screen Z _D =m ² ×Z _conv , the human eye feels a stereo image of an object of interest is a magnified m×A times (x and y direction) and m ² times (z direction) 3D image without distortion.

A kind of stereo image measurement instruction is to establish the two left and right sides of one focus on the focus object based on the geometric relationship between two independent cameras with the same centerline and parallel to each other and an object of interest and the principle of equivalent convergence. The relationship between the parallax of the image and the spatial coordinates of the point of interest in the real scene; establish a relationship between the area of the image on the surface of the object of interest and the actual area of the object of interest in the real scene. A stereo image measurement command can accurately determine the spatial coordinates (x, y, z) of a point of interest depends on whether the left and right images of the point of interest can be accurately positioned in a left-right format image screenshot or left and right two images. _{The abscissas X L} and X _R in the left image screenshot and the right image screenshot in two independent image screenshots. The left and right images of a point of interest are located on the same horizontal line in one left and right format image screenshot or left and right image screenshots in two independent image screenshots, or Y _L = Y _R , where Y _L And Y _R are the ordinates of the focus point in the left image screenshot and the right image screenshot, respectively. In a stereo camera, the left and right images collected by the left and right cameras have parallax in the horizontal direction, and there is no parallax in the vertical direction. The horizontal parallax of the left and right images of a point of interest is P=(X _R −X _L ), and the vertical parallax is zero V=(Y _R −Y _L )=0. The origins of the left and right coordinate systems in a left and right format image screenshot or two independent left and right image screenshots are located at the center of the left image screenshot and the right image screenshot respectively. The coordinate symbols are defined as: X _L and X _R are positive when they are located on the right half of the center vertical axis of the left and right coordinate systems, and negative when they are located on the left half of the center vertical axis of the left and right coordinate systems, respectively. Zero when the center of the coordinate system is on.

For images in a shift-left-right format and a traditional left-right format, the parallax between the left and right images of a point of interest in the real scene in an image screenshot of the left-right format is P=(X _R －X _L ), and the space position of the point of interest Mark is

x=t×(X _L +T/2)÷[T－(X _R －X _L )]－t/2

y=Y _L ÷(m×A)=Y _R ÷(m×A)

z=(A×F×t)÷[T－(X _R －X _L )]

For the left and right independent shift images and the traditional left and right images, the parallax of the left and right images of a focus in the real scene in the left and right independent image screenshots is P=(X _R －X _L ), pay attention The spatial coordinates of the point are;

x=t×(X _L +T/2)÷[T－(X _R －X _L )]－t/2

y=Y _L ÷(m×A)=Y _R ÷(m×A)

z=(A×F×t)÷[T－(X _R －X _L )]

In the following description of the measurement process and method of a stereo image measurement instruction, only the positioning and measurement process and method of the left and right images of a focus point in an image screenshot of the left and right format are taken as an example. The positioning and measurement process and method of the left and right images of a point of interest in the left and right independent image screenshots are exactly the same as the positioning and measurement processes and methods in the left and right format image screenshots.

A three-dimensional image measurement instruction determines the spatial coordinates (x, y, z) of a point of interest based on the left and right images of a point of interest on an object of interest; the first step is to obtain a left and right image that includes the point of interest. An image screenshot of the left and right format of the image; the second step, use the touch screen pen to click and confirm the abscissa X _L of the left image of the focus point on the screen in the left image screenshot; the third step, when the left image of the focus point is on the left When the position in the image screenshot is located on a reference image with geometric characteristics, for example, a non-horizontal straight line, a curve, a geometric mutation on the surface of the object or a geometric feature, the right image of the point of interest is in the right image screenshot The abscissa X _{R of} is located on a _{horizontal straight line that passes through X L} and crosses the left and right image screenshots, at the intersection of the left image of the point of interest with the reference object image with the same geometric characteristics in the left image screenshot. _{Use the touch screen pen to click and determine the abscissa X R} of the right image of the focus point in the right image screenshot. _{After the horizontal coordinates X L} and X _R of the left and right images of a point of interest are located in a screenshot of the left and right format image, the parallax of the two images of the point of interest is P = (X _R －X _L ) and the spatial coordinates (x ,y,z) is determined.

A stereo image measurement process starts with the following two steps. The first step is to obtain an image screenshot in the left and right format including one or more points of interest on the surface of the object of interest, the surface of interest, the volume of interest, surface cracks or the uneven part of the damaged surface from the image; the second step, in the menu Select the destination of this measurement (not limited to), point-camera, point-point, point-line, point-plane, surface area, volume, surface crack, surface crack area, surface crack cross section, surface damage parameter, surface Damaged area, damaged surface cross-section and maximum depth.

The measurement process and method of the distance between a focus point a and the camera lens: the first step is to obtain a screenshot of the left and right format image from the image; the second step, select "point-camera" in the menu; the third step, use the touch screen _{Click the pen to determine the horizontal coordinate X La} of the left image of the focus point a in the left image screenshot, and a horizontal _{line passing through the X La} coordinate and across the left and right image screenshots will automatically appear on the screen; the fourth step, use the touch screen The pen clicks on the horizontal line of the right image screenshot and determines the abscissa X _Ra of the right image of the focus point a in the right image screenshot. The distance from a focus a to the camera is;

Dc=√[xa ² +ya ² +(za－c) ² ]

Among them, c is the distance from the center of the camera to the center of the outer surface of the objective lens.

The measurement process and method of the distance between the two focus points a and b: the first step is to obtain a screenshot of the left and right format from the image; the second step is to select "point-point" in the menu; the third step is to determine the two respectively _{The horizontal coordinates X La} , X _Ra , X _Lb and X _Rb of the left and right images of the two focus points a and b in the left and right image screenshots. The distance between the two attention points a and b is;

Dab＝√[(xb－xa) ² +(yb－ya) ² +(zb－za) ² ]

The measurement process and method of the distance from a point of interest a to a straight line in space: The first step is to obtain a screenshot of the left and right format from the image; the second step is to select "point-line" in the menu; the third step is to determine separately _{Focus on the horizontal coordinates X La} and X _Ra of the left and right images of point a in the screenshots of the left and right images; the fourth step is to determine the two feature points b and c on a straight line in the space. The left and right images are on the left and right. _{The abscissas X Lb} , X _Rb , X _Lc and X _Rc in the two image screenshots. The distance from a focus point a to a straight line passing through two feature points b and c is:

Da-bc=√{[xa－λ(xc－xb)－xb] ² +[ya－λ(yc－yb)－yb] ² +[za－λ(zc－zb)－zb)] ² }

Among them, λ=[(xb－xa)×(xc－xb)+(yb－ya)×(yc－yb)+(zb－za)×(zc－zb)]÷[(xc－xb) ² + (yc－yb) ² +(zc－zb) ² ]

The measurement process and method of the distance from a point of interest a to a spatial plane: The first step is to obtain a screenshot of the left and right format image from the image; the second step, select "point-plane" in the menu; the third step, determine separately _{Focus on the horizontal coordinates X La} and X _Ra of the left and right images of point a in the left and right image screenshots; the fourth step is to determine the three feature points b, c, and d that are located on a spatial plane but not on a straight line. The horizontal coordinates of the left and right images are X _Lb , X _Rb , X _Lc , X _Rc , X _Ld and X _Rd in the left and right image screenshots. The distance from a point of interest a to a plane that includes three feature points b, c, and d that are not on a straight line is;

Da-(bcd)=[I Axa+Bya+Cza+D I]÷√(A ² +B ² +C ² )

Among them, A, B, C are obtained from the following determinant, D =-(Axb+Byb+Czb)

Moving the stylus on the touch screen, the three different paths of the finger or the mouse from one pixel to the next adjacent pixel are along the horizontal direction, the vertical direction and a triangle with horizontal and vertical pixels as right-angled sides. The hypotenuse direction. A curve on the touch screen can be roughly regarded as a horizontal straight line between a large number of adjacent pixels, and the horizontal and vertical lines between the vertical straight line and the adjacent two pixels are a right-angled triangle oblique. A splicing curve formed by splicing edges. The larger the resolution PPI of the touch screen, the closer the actual length of a curve to the length of a splicing curve. Similarly, the area enclosed by a closed-loop curve is closer to the total area of all pixel units enclosed by a closed-loop splicing curve. The horizontal distance between two adjacent pixels is a and the vertical distance is b. The sum of the area of all pixels surrounded by a closed-loop splicing curve is Ω=∑(a×b)+∑(a×b)÷2. The actual surface area of the object of interest is Q=Ω÷(m ² ×A×B).

A measurement process and method focusing on surface area: The first step is to obtain a screenshot of the left and right format from the image; the second step, select "Area" from the menu, the system will automatically retain one of the image screenshots and zoom in to the entire Screen; The third step is to use a touch screen pen to draw a closed-loop splicing curve along the edge of the image of the surface of interest on the screen. The image area enclosed by the closed-loop splicing curve is the area of the image of the surface of interest. The surface area of interest is the area of the image of the surface of interest divided by (m ² ×A×B).

The area of the surface of interest described in the above [0042] is only the area where the actual area of the surface of interest is projected on a plane perpendicular to the center line (Z axis) of the stereo camera. The fourth step is to return to the image screenshots in the left and right format. When the surface of the object of interest is a flat surface or a curved surface with a radius of curvature that is much larger than the length of the surface, determine the planes respectively according to the method described in [0040] above On the surface, the left and right images of the three feature points b, c and d that are not on the same straight line have the abscissas X _Lb , X _Rb , X _Lc , X _Rc , X _Ld and X _Rd in the left and right image screenshots. The actual area of a surface of interest is equal to the surface area of interest obtained by the method described in [0042] above divided by the cosine of the angle between the normal vector N of the surface of the object of interest and the center line (Z axis) of the stereo camera.

A process and method for measuring the volume of a flat panel: the first step is to obtain a screenshot of the left and right format from the image; the second step is to select the volume in the menu; the third step is according to the method described in [0043] above Obtain the actual area of the surface of the attention plate; the fourth step, when the attention plate is a curved surface with a radius of curvature much larger than the length of the surface, determine the left and right sides of the two characteristic points a and b with typical thickness on the attention plate. _{The abscissas X La} , X _Ra , X _Lb and X _{Rb of} each image in the left and right image screenshots. The thickness of a plate of interest is equal to the distance between two characteristic points a and b multiplied by the cosine of the angle between the vector ab and the normal vector N of the surface of the plate of interest. The actual volume of a slab of interest is equal to the actual area of the slab of interest obtained in the third step above multiplied by the thickness of the slab obtained in the fourth step.

The measurement process and method of the cross section of an object surface crack: The first step is to adjust the position and direction of the center line of the stereo camera to be consistent with the longitudinal direction of the crack and parallel to the surface of the object. When the cross section of the crack of interest is opened, a screenshot of the left and right format is taken; the second step is to use the touch screen pen to determine the two intersection points a and b at the two edges of the surface of the object of interest and the crack cross section opening. The horizontal coordinates of the image in the left and right image screenshots are X _La , X _Ra , X _Lb and X _Rb ; in the third step, select "Crack Cross Section" in the menu, and the system will automatically retain one of the image screenshots and zoom in to the entire screen. _{Use a touch screen pen to determine the abscissas X L1} , X _L2 , X _L3 , ... and X _R1 , X _R2 , of multiple characteristic points with inflection points, turning points and peak points on the left and right edges of the crack cross-sectional opening. X _R3 ,……. There is no relationship between _{the characteristic point X L#} on the left edge of the _{crack opening and the characteristic point X R#} on the right edge of the crack opening. The abscissa of each feature point X _L# and X _R# and the above two intersection points a and b are on the same crack cross section, and the parallax of the feature points on the left and right opening edges of all crack cross sections and point a The parallax is the same as the point b, or the convergence depth coordinate Zc of the point a and the point b is the same as the stereo image depth coordinate Zc of all the characteristic points on the left and right crack opening edges of the crack cross section. The left edge of the opening of the crack cross-section is composed of a straight line _{starting from point a, which sequentially connects all adjacent characteristic points X L# on the left edge of the crack cross-section opening.} The right edge of the opening of the crack cross section is composed of a straight line _{starting from point b, which successively connects all adjacent characteristic points X R# on the right edge of the crack cross section opening.} The left and right edges of the crack cross-section form a "V"-shaped cross-sectional opening. The more feature points are selected, the closer the edge of the crack cross-section is to the edge of the actual crack cross-section. The vertical distance Y _L# between point a and each feature point X L# on the left edge of the crack cross-sectional opening _{Y L#} and the vertical distance Y _{R between point b and each feature point X R#} on the right edge of the crack cross-sectional opening _# , The distance between point a and point b or the width of the crack cross section are listed on the cross section diagram.

The measurement process and method of the cross-section and maximum depth of the concave-convex part of an object surface: Here, only the depression caused by the damage or corrosion of the surface of the object is described as an example. The first step is to adjust the position and direction of the center line of the stereo camera to be parallel to the surface of the object. When the typical features and interesting parts of the surface depression of the object are seen on the screen, a left-right format image screenshot is taken, and one of the image screenshots is retained. _{And zoom in to the full screen; the second step is to determine the horizontal coordinates X La} , X _Ra , X of the left and right images of the two intersection points a and b where the surface of the object and the edge of the damaged cross section intersect in the left and right image screenshots. _Lb and X _Rb ; the third step, select "damaged cross section" in the menu, and enter the radius of curvature of the damaged surface +R, (convex surface) or -R (concave surface) in the next level of the menu. A curve with a radius of curvature of R passing through point a and point b will appear on the screen. If the radius of curvature of the damaged surface cannot be obtained, use a touch screen pen to draw a splicing curve between the two intersection points a and b. The splicing curve is smoothly linked with the left surface curve of point a and the right surface curve of point b. In the fourth step, use a touch screen pen to draw a splicing curve between the two intersection points a and b along the edge of the damaged part in the cross-sectional view. The closed-loop splicing curve of the damaged cross-section is composed of a curve with curvature R between points a and b and a splicing curve. Step 5: Go back to the left and right image screenshots, click on the stitching curve and determine the abscissa X _Lc and X _{Rc of the} lowest point C of the damaged section. The area of the damaged cross section of an object surface, the distance between point a and point b, and the vertical distance Yc from the lowest point c of the cross section are all listed on the cross section diagram.

In the actual measurement process, when the measurement purpose and requirements are different from the above-mentioned basic measurement methods, different and reasonable measurement methods and solutions need to be proposed according to different situations. The new measurement method and solution can be a combination of the above-mentioned basic measurement methods or other new methods.

A stereo image positioning and tracking command is based on the principle of equivalent convergence, a focus point or a line of focus on the left and right of the left and right images captured by two lens groups or cameras that are independent of each other, the same and the center line is parallel to each other. After the position of the image or the right image in a left and right format image screenshot or two separate left and right image screenshots in the left image screenshot or right image screenshot is located, locate and track the point of interest or focus on the straight right image or left image The position in the right image screenshot or the left image screenshot in the same left and right format image screenshot or two independent left and right image screenshots. A three-dimensional image positioning and tracking instruction includes three different processes: image positioning, image matching, and image tracking. First, the positioning process is to use a rectangular box to enclose a point of interest or a point of interest. The four perimeters of the rectangular box are parallel to the two coordinate axes in the left and right image screenshots. The center of the rectangular box is The point of the same name in the rectangular box. The positioning process is to determine the positions of the points with the same name of the rectangular box in the left and right image screenshots respectively. The rectangular box surrounding a point of interest is a square box, and the point of interest is also a point with the same name as the square box. The rectangular box surrounding a line of interest is a rectangular box. The center of the rectangular box is the midpoint of the attention line or the point of the same name, and a diagonal line of the rectangular box is the attention line. Second, the matching process is a process of feature matching combined with a simplified gray-scale matching instruction, which is a process of feature and gray-scale search, contrast, comparison and matching of images limited to a limited rectangular box. The matching content includes the relationship between the left and right images and the reference object, corner points, edge points, edge lines, and other geometric features, and the color features, surface textures, color and texture change patterns and rules in the rectangular box. Third, the tracking process is that when a point of interest or the left and right images of a line of interest are located, when the point of interest or the image of the line of interest moves to a new position, the automatic tracking has been positioned and surrounded by a rectangular box. The new position, coordinates, parallax and distance to the stereo camera of the point with the same name in the rectangular box at any moment in the left and right images of the point of interest or the line of interest respectively in the left and right image screenshots. The reason for the movement of a point of interest or an image of a line of interest can be a change in the position of the point of interest or the line of interest and a change in the position or angle of the stereo camera.

The image positioning process of a point of interest or a line of interest: the first step, for a point of interest a, use a touch screen pen to click on the screen at the left image of the point of interest a. A square box encloses the focus point a, and the center of the square box is the left image of focus point a, or the point with the same name, with coordinates (X _La , Y _La ). For a focus on straight line bc, using the stylus along the screen to slide the other endpoint b c bc straight left video image from a left end of the line bc. A rectangular box encloses the left image of the attention line, and the center of the rectangular box is the midpoint or the point of the same name of the left image of the attention line bc. The left image of the attention line bc is a diagonal line of the rectangular box. The coordinates of the two end points b and point c of the left image of the attention line bc _{are (X Lb} , Y _Lb ) and (X _LC , Y _LC ), respectively. In the second step, the matching process starts to search for and locate the same feature in the right image screenshot with the same name as the left image in the left image screenshot. The points with the same name have the following features in the left and right screenshots; the first feature is a point of interest or a line of interest on the left image in the left image screenshot on the reference object, corner points, edge points, edge lines and others For geometric features, the point with the same name in the screenshot on the right is also located on the same geometric feature's reference, corner point, edge point, edge line and the same geometric feature; the second feature is a point of interest and a line of interest The positions of the points with the same name in the left and right image screenshots are located on a horizontal line that crosses the left and right image screenshots; the third feature is that the ordinates of the two end points b and point c of a line of interest are equal, Y _Lb = Y _LC ; The fourth feature is that the color, surface texture, color and texture change patterns and laws in a rectangular box enclosing a point of interest or a line of interest are consistent; the fifth feature is the pattern and feature matching, The search, comparison, and comparison process is limited to a limited rectangular box. After the matching is completed, confirm that the coordinates of the point of interest and the right image of the line of interest in the right image screenshot are (X _Ra , Y _Ra ), (X _Rb , Y _Rb ) and (X _RC , Y _RC ), with the same name The parallaxes corresponding to the points are (X _Ra －X _La ) and (X _Rbc －X _Lbc ).

The necessary and sufficient condition for the coincidence of the focal plane of the eye and the image plane of the stereoscopic image in the above [0028] is A=[T÷(F×t)]×Z=k×Z, and the visual effect of the same-screen principle Made a qualitative explanation. According to the above

How to obtain the parallax and process of two images of a focus, here is a quantitative calculation of the principle of the same screen, to obtain the quantitative result of the need to change the screen magnification. When the stereo depth Z of an object of interest in the real scene changes, a stereo image positioning tracking command will automatically track the change of the position of the object of interest, and bring the change of the disparity of the point of the same name into the formula (1) to obtain;

ΔZ=(A×F×t)÷{[T－(X _R2 －X _L2 )] ^-1 －[T－(X _R1 －X _L1 )] ^-1 }. Substitute the ΔZ result of formula (1) into formula (2) to obtain; ΔA=[T÷(F×t)]×ΔZ=[T÷(F×t)]×ΔZ=(A×T)×{[ T－(X _R2 －X _L2 )] ^-1 －[T－(X _R1 －X _L1 )] ^-1 }. In formula (1), the screen magnification A (=W/w) is a constant. In formula (2), the screen magnification ratio A comes from formula (1) and has nothing to do with ΔA. According to the principle of the same screen, the enlargement and reduction of the stereoscopic image in the stereoscopic playback screen will not have any influence on the position of a focus in the real scene. This is the meaning of "equivalent" in the equivalent change, equivalent process, and equivalent result described in [0028] above. A three-dimensional image on-screen instruction makes the ΔA obtained by the formula of the image played on the screen synchronously change with the three-dimensional depth Z of an object of interest in the real scene. At this time, the convergence points of the left and right images of the object of interest will be directly implemented on the screen, and the position of the focal plane of the eye and the position of the image plane of the stereoscopic image will overlap. The distance between an object of interest and the stereo camera can be measured in real time by an external laser or infrared rangefinder, or by a same-screen chip built in the image processor. Compared with peripheral devices, a same-screen chip has the advantages of faster speed, higher efficiency, smaller delay, more convenient operation, smaller size, lower cost and more user-friendly.

An equivalent convergence point reset command is to set the object as a new focus object through the stereo image of an object on the screen during the playback of the stereo image, and then use the stereo image of the new focus object to converge the equivalent of the stereo camera The point is reset to the new object of interest. _{According to the formula Z conv} =(A×F×t)÷T described in the above [0029], changing the screen magnification A can change the position Z _{conv of the} equivalent convergence point M of an object of interest. In fact, an equivalent convergence point reset instruction combined with other instructions perfectly solves the three current application requirements and problems. The first application is that the stereo player can become a healthy stereo player; the second application is that the audience can interact with the content being played in a stereo player; the third application is during shooting, the stereo camera lens shoots When the subject transfers from one object of interest to another new object of interest, the equivalent convergence point of a core-moving stereo camera needs to be transferred from the previously set object of interest to the new object of interest.

The definition of a healthy 3D player is a 3D player in which the convergent point of the 3D image of the object of interest in the 3D image played in the 3D player appears on the screen. First of all, by setting a same-screen chip in a stereo player, most stereo players can be turned into healthy stereo players. Secondly, after setting up a same-screen chip in a stereo player, the audience can have a new kind of interventional interaction with the content being played in the stereo player, feel and participate in it. First, there are multiple images of different characters or objects of interest surrounded by boxes on the screen, and the audience uses the remote control to determine which one of them is most interested in a new object or character of interest. In fact, the image of the new object of interest determined by the audience on the screen is a three-dimensional image of the convergence of the left and right images of the new object of interest. Secondly, a same-screen chip will obtain an image screenshot from the input image in a left and right format or two independent images, and determine the left and right images according to the process and method described in [0048] and [0049]. The coordinates of the points with the same name in the two image screenshots surrounding the left and right new object images of interest, so as to obtain the disparity of the two points with the same name on the left and right P=(X _R －X _L ) and enter the formula Z=(A×F× t)÷[T－(X _R －X _L )] and the formula Z _C =Z _D ×[T÷(A×F×t)]×(A×F×t)÷[T－(X _R －X _L )]=(Z _D ×T)÷[T－(X _R －X _L )], Z _{C is} obtained. Let the obtained Z _C =Z _conv =(A×F×t)÷T, determine the position of the equivalent convergence point M or the stereo depth Z _{conv of} the new object of interest and the distance to be moved h=(F×t) ÷(2Z _conv ). Summarize the above process; firstly, a same-screen chip set in a stereo player obtains a left-right format or two independent image screenshots from the input stereo image, and captures the left and right images of a newly determined object of interest. Perform positioning, matching and tracking, and obtain the stereo depth Z _{C of the} convergence point of the stereo image of the new object of interest, let Z _C =Z _conv ; The second step is to correct the amount of core shift of the object of interest h. If the content being played comes from a stereo camera designed according to the principle of equivalent convergence, the core shift amount h represents the correction of the core shift amount of a newly set new object of interest. If the content comes from a stereo camera that uses the parallel method to shoot, the core shift amount h represents the change of the stereo camera to a stereo camera that satisfies the principle of equivalent convergence. If the content comes from a stereo camera that uses the convergence method to shoot, the focus plane of the eye and the image plane of the stereo image still cannot perfectly coincide. In the third step, a same-screen chip locates, matches and tracks the left and right images of the new object of interest through the processes and methods described above, including the position of the point with the same name, coordinates, parallax and the distance to the stereo camera, in real time Change the screen magnification rate and ensure that the convergence point of the stereo image of the new object of interest is implemented on the stereo player screen.

The above-mentioned basic measurement method appears inconvenient, lacks efficiency, and is not easy to accurately determine the position of the right image of a point of interest in the right image screenshot. A same-screen chip simplifies the above-mentioned basic measurement process to one or two steps to accurately locate the position of the right image of a point of interest in the right image screenshot, making the real-time measurement process of the stereo image simpler, more efficient, and more efficient. Humane and precise. At the same time, line/diameter/height, graphic matching, and volume are added to the menu.

The measurement process and method of a same-screen chip are as follows: First, manually determine the abscissa X _L of the left image of a point of interest in the left image screenshot in a left-right image screenshot. A same-screen chip will match the left and right images of the point of interest around the same feature at the point with the same name, obtain the abscissa X _{R of the} point with the same name in the right image screenshot, and calculate the parallax of the point of interest P = (X _R － X _L ) and measurement results.

The measurement process and method of the distance between a focus point a and the camera lens: The first step is to obtain an image screenshot in the left and right format from the image, save one of the image screenshots and zoom in to the full screen; in the second step, select "Point" in the menu -Camera"; the third step, use the touch screen pen to click and determine the position of point a. A chip on the same screen will calculate the distance from a focus point a to the midpoint of the line connecting the midpoints on the outer surfaces of the two camera objective lenses as;

Dc=√[xa ² +ya ² +(za－c) ² ]

The measurement process and method of the straight-line distance between the two focus points a and b: The first step is to obtain an image screenshot in the left and right format from the image, save one of the image screenshots and zoom in to the full screen; the second step, the menu Select "straight line/diameter/height"; in the third step, use the touch screen pen to click and confirm the position of point a and keep the touch screen pen sliding to the position of point b on the screen. A chip with the same screen will calculate the distance between the two focus points a and b as:

Dab＝√[(xb－xa) ² +(yb－ya) ² +(zb－za) ² ]

The process and method of measuring the distance from a point of interest a to a straight line in space: the first step is to obtain an image screenshot in the left and right format from the image, save one of the image screenshots and zoom in to the full screen; the second step, select from the menu "Point-line"; the third step, use the touch screen pen to click and confirm the position of point a; the fourth step, use the touch screen pen to click and confirm the position of point b in a straight line and keep the touch screen pen on the screen Slide to point c. A chip on the same screen will calculate the distance from a focus point a to a straight line passing through two feature points b and c;

Da-bc=√{[xa－λ(xc－xb)－xb] ² +[ya－λ(yc－yb)－yb] ² +[za－λ(zc－zb)－zb)] ² }

The measurement process and method of the distance from a point of interest a to a spatial plane: The first step is to obtain an image screenshot in the left and right format from the image, save one of the image screenshots and zoom in to the full screen; the second step, select from the menu "Point-plane"; the third step, use the touch screen pen to click and confirm the position of point a; the fourth step, use the screen pen to click and confirm the position of point b and keep the touch screen pen continuously sliding on the screen to points c and The position of point d, where point b, point c, and point d are three points that are not all on a straight line. A chip with the same screen will calculate the plane distance from a focus a to a three feature points b, c and d that are not all in a straight line;

Da-(bcd)=[I Axa+Bya+Cza+D I]÷√(A ² +B ² +C ² )

A same-screen chip can not only be applied to the axis-shifting stereo camera, but also can be applied to all two independent, identical, and parallel-centered stereo cameras, and make the stereo images collected by the stereo camera have the equivalent convergence method The three-dimensional image obtained has the same three-dimensional effect.

The axis-shifting stereo camera proposed by the present invention not only solves the problems existing when the current stereo camera collects and plays stereo images, it has a highly integrated structure design, a humanized operation method, and has simple operation and image restoration. High, low image delay, low cost, easy to promote and popularize.

Description of the drawings

Figure 1-1 A schematic diagram of a top view of a dual-prism axis-shifting stereo camera with a single image sensor;

Figure 1-2 A schematic view of A direction view of a dual-prism shifting stereo camera with a single image sensor;

Figure 2-1 A schematic top view of an axis-shifting stereo camera with a single image sensor;

Figure 2-2 A schematic diagram of a direction A view of a single-image sensor axis-shifting stereo camera;

Figure 3-1 A schematic top view of a single-image sensor variable viewing distance axis-shifting stereo camera;

[Corrected according to Rule 26 28.10.2020]
Figure 3-2 A schematic view of the A direction view of a single-image sensor variable viewing distance axis-shifting stereo camera;

Figure 4-1 A schematic diagram of a top view of a dual-image sensor axis-shifting stereo camera;

[Corrected according to Rule 26 28.10.2020]
Figure 4-2 A schematic view of the A direction view of a dual-image sensor axis-shifting stereo camera;

Figure 5-1 Schematic diagram of the relative position of the image sensor and the smallest imaging circle of the axis before shifting;

Figure 5-2 Schematic diagram of the relative position of the image sensor and the smallest imaging circle of the axis after the axis is shifted;

Figure 6-1 A schematic diagram of a three-dimensional image collection space;

Figure 6-2 A schematic diagram of a 3D video playback space;

Figure 7-1 Schematic diagram of the shooting principle of the three-dimensional image convergence method;

Figure 7-2 Schematic diagram of the principle of parallel shooting of stereo images;

Figure 7-3 Schematic diagram of the shooting principle of the stereo image equivalent convergence method;

Figure 8 Schematic diagram of equivalent convergence method and parallax principle;

Figure 9-1 Schematic diagram of the image plane on the screen;

Figure 9-2 The image plane is in front of the focal plane;

Figure 9-3 The image plane is behind the focal plane;

Figure 9-4 Schematic diagram of the principle that the image plane and the focal plane are on the same screen;

Figure 10 A schematic diagram of the positions of the left and right images of a point of interest in a left and right format screenshot;

Figure 11 A schematic diagram of the coordinate of any point in space and the parallax principle of the image sensor after the axis is shifted;

Figure 12 A schematic diagram of measuring the distance from a point of interest to a stereo camera;

Figure 13 Schematic diagram of measuring the distance between two points of interest;

Figure 14 Schematic diagram of measuring the distance from a point of interest to a straight line;

Figure 15 Schematic diagram of measuring the distance from a point of interest to a plane;

Figure 16 A schematic diagram of measuring the surface area of a flat object;

Figure 17 Schematic diagram of measuring the volume of a flat object;

Figure 18-1 Collecting a schematic diagram of the transverse interface of a surface crack;

Figure 18-2 Schematic diagram of measuring a cross-section of a surface crack;

Figure 19-1 A schematic diagram of a cross-sectional view of a damaged depression on the surface;

Figure 19-2 Schematic diagram of measuring a cross-section of a damaged surface.

Detailed ways:

The specific embodiment of the present invention represents an example of the embodiment of the present invention, and has a corresponding relationship with the content and specific matters in the claims and the specification. The present invention does not limit the embodiments, and can be embodied in various different embodiments within the scope not departing from the gist of the present invention. The illustrative cases in all the schematic diagrams are examples of the multiple technical solutions that can be implemented.

Figure 1-1 shows a schematic top view of a dual-prism shifting stereo camera with a single image sensor. In the figure, there are two independent and identical lens groups (a) and (b) on the left and right. After shifting, the two lens groups (a) and (b) are simultaneously shifted toward each other by a distance of L, and the center lines of the two lens groups (a) and (b) are shifted from the position 7 before shifting To the position of 8, the distance between each other is t. t is the viewing distance of the tilt-shift stereo camera. A partition plate 2 is arranged on the center line of the tilt-shift stereo camera. The two right-angled prisms 3 have a right-angled triangle-shaped surface 6 coated with a coating and bonded together.

Figure 1-2 shows a schematic view from the direction of A with a single image sensor and double prism shifting stereo camera. In the figure, in lens group (b), the image collected by lens group 1 enters the right-angle prism 3 through the right-angle incident plane of a right-angle prism 3 and is bent down by 90° by the total reflection of the inclined surface, and is projected to an image sensor 5 for effective imaging Image on the left half of surface 4.

Figure 2-1 shows a schematic top view of a single-image sensor pivot-shifting stereo camera. In the figure, there are two independent and identical lens groups (a) and (b) on the left and right. After shifting, the two lens groups (a) and (b) are simultaneously shifted toward each other by a distance of L, and the center lines of the two lens groups (a) and (b) are shifted from the position 7 before shifting To the position of 8, the distance between each other is t. t is the viewing distance of the tilt-shift stereo camera. A partition plate 2 is arranged on the center line of the tilt-shift stereo camera.

Figure 2-2 shows a schematic view of a single-image sensor axis-shifting stereo camera A in the direction of view. In the figure, in the lens group (b), the image collected by the lens group 1 is directly projected onto the left half of the effective imaging surface 4 of an image sensor 5 for imaging.

Figure 3-1 shows a schematic top view of a single-image sensor variable viewing distance axis-shifting stereo camera. In the figure, a pair of turning prisms 9 and 10 are provided in the left and right independent and identical lens groups (a) and (b). A pair of turning prisms 9 and 10 divide the lens group (a) or (b) into the front objective lens group 1 and the rear imaging lens group 11. In a lens group, the image collected by the objective lens group 1 enters the front right-angle prism 9 prism and then is bent 90° by the inclined surface of the right-angle prism 9 and enters the right-angle prism 10, and the inclined surface of the right-angle prism 10 again bends the image by 90° into the imaging The lens group 11, the imaging lens group 11 projects an image on the left half or the right half of the effective imaging surface 4 of an image sensor 5 for imaging. After the axis is shifted, the rear right-angle prism 10 and the imaging lens group 11 in the two lens groups (a) and (b) are simultaneously translated by a distance of L in directions opposite to each other. The center lines of the rear right-angle prism 10 and the imaging lens group 11 in the two lens groups (a) and (b) are translated from the position 7 before the shift to the position 8, and the distance between each other is t. During the shifting process, the positions of the objective lens group 1 and the front right-angle prism 9 remain unchanged, and the distance between each other is t'. t'is the viewing distance of the tilt-shift stereo camera. A partition plate 2 is arranged on the center line of the tilt-shift stereo camera.

Figure 3-2 shows a schematic view of a direction A view of a single-image sensor variable viewing distance axis-shifting stereo camera. The A direction view shown in the figure, the imaging path of the lens group (b).

Figure 4-1 shows a schematic top view of a dual-image sensor shift stereo camera. In the figure, there are two independent and identical lens modules (a) and (b) on the left and right. In a lens group, the image collected by the lens group 1 is directly projected onto the effective imaging surface 4 of an image sensor 5 for imaging. After shifting, the two lens groups (a) and (b) are simultaneously translated by a distance of L toward each other, and the center lines of the two lens groups (a) and (b) are respectively 7 positions before shifting Shift to the position of 8, the distance between each other is t. t is the viewing distance of the tilt-shift stereo camera.

Figure 4-2 shows a schematic view of a dual-image sensor axis-shifting stereo camera from the A direction. The A direction view shown in the figure, the imaging path of the lens group (b).

Figure 5-1 shows a schematic diagram of the relative position of the image sensor and the smallest imaging circle of the axis before shifting. In the figure, the effective imaging surface 4 of an image sensor is covered by an imaging circle 12 with a vertical centerline of 7 and a radius of r. The center of the effective imaging surface 4 of the image sensor coincides with the center of the imaging circle 12. The horizontal length of the effective imaging surface 4 of the image sensor is w and the vertical height is v.

Figure 5-2 shows a schematic diagram of the relative position of the image sensor and the minimum imaging circle of the shift axis after the shift axis. When the axis is shifted, the imaging circle 12 is translated by a distance of L in the right direction along the horizontal direction, and the position of the effective imaging surface 4 of the image sensor remains unchanged. After shifting, the horizontal distance between the vertical centerline 8 of the imaging circle 12 at the new position and the vertical centerline 7 of the imaging circle 12 before shifting is L. The minimum diameter of the imaging circle 12 is;

D _min ＝2R＝2√[(w/2+L) ² +(v/2) ² ]

Figure 6-1 shows a schematic diagram of a three-dimensional image collection space. The left and right cameras 13 and 14 rotate inwardly around the center of the camera lens at the same time, until the center lines of the two cameras 13 and 14 converge on an object of interest 17 in the real scene to start shooting. This method of shooting stereo images is called the convergence method. The distance between the lens centers of the left and right cameras 13 and 14 is t. The scene in front of the attention object 17 is referred to as the foreground object 18, and the scene behind is referred to as the back scene 19.

Figure 6-2 shows a schematic diagram of a 3D video playback space. The left and right images captured by the left and right cameras 13 and 14 are simultaneously projected onto a flat screen 22 with a width of W. The horizontal distance between the projection of the left and right images 23 and 24 on the screen is the left and right images 23 And 24 parallax P. When a person’s left eye 20 and right eye 21 can only see the projections of the left and right images 23 and 24 on the screen 22, the human brain merges the projections of the two images 23 and 24 seen by the left and right eyes. Perceived three-dimensional images 25, 26, and 27 of objects 17, 18, and 19 of interest.

According to the geometric relationship shown in Figure 6-2, the following relationship is obtained,

Z _C ＝Z _D ×T÷(T－P) (1)

Among them, Z _C -the distance from the midpoint of the connection between the two eyes to the convergence point of the left and right images on the screen

Z _D -the distance from the midpoint of the connection between the eyes to the screen

T-the distance between the eyes

P-Parallax, the horizontal distance between the projection of the left and right images 23 and 24 on the screen

ΔP=P _max －P _min ＝T×Z _D (1/Z _cnear －1/Z _cfar ) (2)

Among them: P _max -the maximum parallax between the left and right images 23 and 24 on the screen

P _min -the minimum parallax between the left and right images 23 and 24 on the screen

Z _{cnear-the shortest distance between the} eyes to the convergence point of the left and right images 23 and 24, (P<0 negative parallax, audience space)

Z _cfar -the farthest distance between the eyes to the convergence point of the left and right images 23 and 24, (P>0 positive parallax, screen space)

Definition, P _rel =ΔP/W

Among them: P _rel -the parallax change per unit width of the flat screen

W-the horizontal length of the flat screen

Figure 7-1 shows a schematic diagram of the shooting principle of the stereo image convergence method. In the figure, when the left and right cameras 13 and 14 capture an object of interest 28 located on the center line of the stereo camera by the convergence method, the object of interest 28 is imaged at the center of the left and right image sensors 15 and 16.

Figure 7-2 shows a schematic diagram of the principle of parallel shooting. In the figure, the center lines of the left and right cameras 13 and 14 are parallel to each other. When shooting an object of interest 28 on the center line of the stereo camera, the imaging of the object of interest 28 on the left and right image sensors 15 and 16 deviates by two. The center of the image sensors 15 and 16.

Figure 7-3 shows a schematic diagram of the shooting principle of the stereo image equivalent convergence method. In the figure, the center lines of the two cameras 13 and 14 are parallel to each other, and an object of interest 28 located on the center line of the stereo camera is photographed. Before shooting, the lens groups of the left and right cameras 13 and 14 are moved in parallel by a distance of L in the horizontal direction toward the opposite direction to each other. An object of interest 28 located on the center line of the stereo camera is imaged at the center of the left and right image sensors 15 and 16.

Figure 8 shows a schematic diagram of the equivalent convergence method and the principle of parallax. In the figure, after shifting the axis, the left and right cameras 13 and 14 photograph an object of interest 17 in the space.

According to the geometric relationship shown in Figure 8, we get the following relationship,

d=t×F×(1/Z _conv －1/Z)=2L－(t×F)÷Z (3)

Among them, the parallax of a point 17 in d-space on the left and right image sensors

L-the translation of the center line of a camera along the horizontal direction

t-the distance between the center lines of the two cameras after shifting the axis

F-the equivalent focal length of the camera lens

Z-the three-dimensional depth of a point 17 in space

Z _conv -the distance of the equivalent convergence point 28

According to formula (3), the following formula can be derived;

Δd=d _max －d _min ＝t×F×(1/Z _near －1/Z _far ) (4)

Among them: d _max -the maximum parallax of the two images on the left and right image sensors

d _min -the minimum parallax of the two images on the left and right image sensors

Z _near -the three-dimensional depth of the foreground object 18 in space

Z _far -the three-dimensional depth of the back scene in space 19

Definition, d _rel =Δd/w

Among them: d _re -parallax change per unit width of the image sensor

w-the horizontal length of the effective imaging surface of the image sensor

Let, P _rel = d _rel

Inferred: t=[(Z _D ÷A×F)×(1/Z _cnear －1/Z _cfar )÷(1/Z _near －1/Z _far )]×T (5) where: A – screen zoom in Rate W/w

The formula (5) shows that the visual distance between the two cameras is not equal to the distance between human eyes.

Let: P=A×d and formula (3) are substituted into formula (1):

Z _C =(Z _D ×T)÷(T－P)=(Z _D ×T)÷(T－A×d)

＝(Z _D ×T×Z)÷[A×F×t－(2A×L－T)×Z] (6)

Equation (6) shows that the _{relationship between Z C} and Z is not linear. Ideal imaging is any point in the 3D image acquisition space. A straight line and a plane correspond to the only point, a straight line and a plane in the 3D image playback space. The sufficient and necessary condition for ideal imaging is that the relationship between the stereo depth Z of an object of interest in the real scene and the stereo depth Z _{C of the} convergent point of the stereo image of the object of interest is a linear relationship. It can be seen from formula (6) that the necessary and sufficient condition for the linear relationship between _{Z C and Z is:}

(2A)×L－T=0 or L=T÷(2A)

Equation (6) is linearized and simplified into the following equation,

Z _C ＝Z _D ×[T÷(A×F×t)]×Z (7)

Formula (7) shows that the relationship between the stereo depth of an object of interest in the real scene and the stereo depth of the convergence points of the two images of the object of interest is a linear relationship.

Figure 9-1 shows a schematic diagram of the image plane on the screen. In the figure, when the projections of the left and right images 23 and 24 on the screen 20 overlap, the parallax P=0 of the left and right images 23 and 24, and the convergence point of a stereoscopic image 25 after brain fusion is located on the screen 22.

Figure 9-2 shows a schematic diagram of the image plane in front of the focal plane. In the figure, when the projection positions of the left and right images 23 and 24 on the screen 20 cross in opposite directions, the parallax P<0 of the left and right images 23 and 24, the convergence point of a stereo image 26 after the brain fusion appears on the screen and Between the audience.

Figure 9-3 shows the image plane behind the focal plane. In the figure, when the projection positions of the left and right images 23 and 24 on the screen 20 are positively crossed, the parallax of the left and right images 23 and 24 is P>0, and the convergence point of a three-dimensional image 27 after brain fusion appears on the screen. rear.

Figure 9-4 shows a schematic diagram of the principle that the image plane and the focal plane are on the same screen. In the figure, by changing the screen magnification A, the projection positions of the left and right images 23 and 24 on the screen 20 are always kept coincident. The convergence point of a three-dimensional image 25, 26 and 27 after the brain fusion is always kept on the screen 20.

Figure 10 is a schematic diagram of the positions of the left and right images of a point of interest in a left and right format screenshot. In the figure, the abscissa of the left image 31 of a focus point a in the left image screenshot 29 in a left-right format image screenshot is X _{L , and X L} <0 according to the symbol rule. _{The right image 32 of the focus point a has X R} , X _R >0 in the right image screenshot 30 in a left-right format image screenshot. The positions of the left image 31 of the attention point a in the left image screenshot 29 and the right image 32 in the right image screenshot 30 are both located on the same horizontal line 33 across the screen. _{The ordinate Y L} of the left image 31 of the focus point a in the left image screenshot 29 is equal _{to the ordinate Y R of the} right image 32 in the right image screenshot 30. The parallax between the left image 31 and the right image 32 of the attention point a is P=(X _R −X _L ).

For an image in a shift left-right format and a traditional left-right format, the parallax of the left and right images of a focus a on the image screenshots 29 and 30 of the left-right format is P=(X _R －X _L ), which is substituted into the formula ( 1) Obtained;

Z _C =Z _D ×T÷(T－P)=(Z _D ×T)÷[T－(X _R －X _L )] (8a)

Substituting formula (7) into formula (8a), simplified to obtain,

Z=(A×F×t)÷[T－(X _R －X _L )] (9a)

For two independent shifting images and traditional two independent images, the left and right image screenshots are two independent image screenshots. The parallax of the left and right images of a focus a in two independent image screenshots is P=(X _R －X _L ), which is obtained by substituting into formula (1);

Z _C =Z _D ×T÷(T－P)=(Z _D ×T)÷[T－(X _R －X _L )] (8b)

Substitute formula (7) into formula (8b), and get the formula after simplification:

Z=(A×F×t)÷[T－(X _R －X _L )] (9b)

Figure 11 shows a schematic diagram of the coordinate of a point in space and the parallax principle of the image sensor after the axis is shifted. According to the geometric relationship shown in Figure 11, the following relationship is obtained,

d ₁ +L=F×(x+t/2)÷Z; d ₂ －L=F×(x－t/2)÷Z

Get the formula for coordinates x and Z:

x=[Z×(d ₁ +L)÷F]－t/2 (10)

For an image in a shift left-right format and a traditional left-right format, d ₁ =X _L /A, L = T/2A and formula (9a) are introduced into formula (10), and the result is simplified,

x=t×(X _L +T/2)÷[T－(X _R －X _L )]－t/2 (11a)

The spatial coordinate a(x,y,z) of a focus point a is;

x=t×(X _L +T/2)÷[T－(X _R －X _L )]－t/2

y=Y _L ÷(m×A)=Y _R ÷(m×A)

z=(A×F×t)÷[T－(X _R －X _L )]

For two independent shift-axis images and traditional two independent images, d ₁ =X _L /A, L = T/2A and formula (9b) are introduced into formula (10) to simplify the result;

x=t×(X _L +T/2)÷[T－(X _R －X _L )]－t/2 (11b)

The spatial coordinate a(x,y,z) of a focus point a is;

x=t×(X _L +T/2)÷[T－(X _R －X _L )]－t/2

y=Y _L ÷(m×A)=Y _R ÷(m×A)

z=(A×F×t)÷[T－(X _R －X _L )]

Figure 12 shows a schematic diagram of measuring the distance from a point of interest to a stereo camera. _{According to the process and method described in the above-mentioned [0081], determine the horizontal coordinates X La} and X _Ra of the left and right images 31 and 32 of a focus point a in the left and right image screenshots 29 and 30, respectively. The distance from a point of interest a to the midpoint of the line connecting the outer surfaces of the objective lenses of the stereo cameras 13 and 14 is:

Dc=√[xa ² +ya ² +(za－c) ² ]

Among them, c is the distance from the center of the lens group of the camera 13 or 14 to the center of the objective lens surface.

Figure 13 shows a schematic diagram of measuring the distance between two points of interest. _{According to the process and method described in [0081] above, determine the horizontal coordinates X La} and X _Ra of the left and right images 31 and 32 of the two focus points a and b in the left and right image screenshots 29 and 30, respectively, X _Lb and X _Rb . The distance between the two attention points a and b is;

Dab＝√[(xb－xa) ² +(yb－ya) ² +(zb－za) ² ]

Figure 14 shows a schematic diagram of measuring the distance from a point of interest to a straight line passing through two feature points. _{The first step is to determine the horizontal coordinates X La} and X _Ra of the left and right images 31 and 32 of the left and right image screenshots 29 and 30 of a focus point a according to the process and method described in [0081] above. . _{The second step is to determine the abscissas X Lb} , X _Rb , X _Lc and X _Rc of the left and right images 31 and 32 of the two feature points b and c on a straight line in the left and right image screenshots 29 and 30 respectively. . The distance from a concern a to a straight line passing through two feature points b and c is;

Da- bc ＝√{[xa－λ(xc－xb)－xb] ² +[ya－λ(yc－yb)－yb] ² +[za－λ(zc－zb)－zb)] ² }

Among them, λ=[(xb－xa)×(xc－xb)+(yb－ya)×(yc－yb)+(zb－za)×(zc－zb)]÷[(xc－xb) ² + (yc－yb) ² +(zc－zb) ² ]

Figure 15 shows a schematic diagram of measuring the distance from a point of interest to a plane. _{The first step is to determine the horizontal coordinates X La} and X _Ra of the left and right images 31 and 32 of the left and right image screenshots 29 and 30 of a focus point a according to the process and method described in [0081] above. . _{The second step is to determine on the plane 34 the horizontal coordinates X Lb} of the left and right images 31 and 32 of the three feature points b, c, and d that are not all on the same straight line. , X _Rb , X _Lc , X _Rc , X _Ld and X _Rd . The distance from a point of interest a to a plane 34 that includes three feature points b, c, and d is:

Da-(bcd)=[I Axa+Bya+Cza+D I]÷√(A ² +B ² +C ² )

Among them, A, B, C are obtained from the following determinant, D =-(Axb+Byb+Czb)

Figure 16 shows a schematic diagram of measuring the surface area of a flat object. A method and steps for measuring the surface area of the plane of interest 36 surrounded by a closed-loop curve 35; the first step is to follow the procedures and methods described in [0041] and [0042] above, using a touch screen pen on the touch screen Draw a closed-loop curve 35 that includes the surface area of a plane of interest 36. The area enclosed by a closed loop curve 35 is obtained. In the second step, according to the process and method described in [0039], the left and right images including the three feature points b, c and d that are not all in a straight line on the surface of the plane of interest 36 are determined respectively. The abscissas in the left and right image screenshots are X _Lb , X _Rb , X _Lc , X _Rc , X _Ld and X _Rd . The actual area of the surface of a plane of interest 36 is equal to the orthographic projection area obtained in the first step divided by a normal vector N determined by the three feature points b, c, and d on the surface of the plane of interest 36 sandwiched between the Z axis The cosine of the angle.

Figure 17 shows a schematic diagram of measuring the volume of a flat object. A method and procedure for measuring the volume of the plate of interest; the first step is to obtain the actual area of the surface 38 of the plate of interest 37 according to the process and method described in [0087] above. In the second step, according to the process and method described in [0043] above, the actual thickness at the two characteristic points a and b with thickness on the attention plate 37 is obtained equal to the length of the two characteristic points a and b multiplied by two The cosine of the angle between the vector ab formed by the feature points and the surface normal vector N of the attention plate 37. The actual volume of a plate 37 of interest is equal to the actual area of the surface 38 of the plate 37 multiplied by the actual thickness.

Figure 18-1 shows a schematic diagram of a cross-section of a surface crack. In the figure, a crack 39 appears on the surface of an object of interest. Method and steps for measuring the shape and depth of the opening at the cross section 40 of the surface crack: According to the process and method described in [0045] above, the first step is to adjust the centerline of the stereo camera to be consistent with the longitudinal direction of the crack 39 and to be consistent with the surface of the object. parallel. When a representative position in the cross section 40 of the crack on the surface of the object is seen on the screen, an image screenshot 29 and 30 in the left and right format is collected.

Figure 18-2 shows a schematic diagram of measuring a cross-section of a surface crack. The second step is to determine the distance V between the left and right edges of the crack 39 at the crack cross section 40 and the two intersection points a and b of the surface of the object of interest, where V is the surface crack width of the crack 39 at the crack cross section 40. The third step is to use a touch screen pen, finger or mouse to determine the characteristic points X _L1 , X _L2 , X _{L3 , ... on the left edge of the crack 39 and the characteristic points X R1} , X _R2 , X _R3 , on the right edge, respectively. …….. The left and right edges of the crack 39 are composed of straight line segments _{connecting the adjacent feature points X L#} and X _R# on the left and right edges of the crack 39 respectively, starting from point a and point b respectively. _{The vertical heights y L#} and y _R# between each feature point X _L# and X _R# and point a and point b respectively represent the depth of the feature point from the surface of the object of interest.

Figure 19-1 shows a schematic diagram of a cross-sectional view of a damaged depression on the surface. In the figure, a recessed portion 41 appears on the surface of an object of interest. Method and steps for measuring the cross section 42 of the recessed part of the surface of the object: The first step is to adjust the center line of the stereo camera to be parallel to the surface of the recessed part of the object and collect one when a representative part of the recessed part 41 of the object surface is seen on the touch screen. Image screenshots 29 and 30 in left and right format.

Figure 19-2 shows a schematic diagram of a cross-sectional view of a damaged depression on the surface. The second step is to determine the distance U between the two intersection points a and b of the cross section 42 and the surface of the object. The third step is to select "damaged cross section" in the menu of the touch screen and enter the radius of curvature of the surface of the object at the cross section of the damaged part +R (convex surface) or -R (concave surface). A curve 43 passing through points a and b and a radius of curvature R will appear on the touch screen. The fourth step is to use a touch screen pen, finger or mouse to draw a curve 44 between the two intersection points a and b along the edge of the recessed part in the image screenshot. A closed loop curve on a concave cross section 42 on the surface of the object is composed of a curve 43 with a radius of curvature of R and a curve 44 on the edge of the image of the concave portion. The fifth step is to determine the position of the lowest point c of the cross section 42 in an image screenshot. The depths ya and yb between the point a and the point b from the point c and the area of the cross section 42 (shaded part in the figure).

Claims (2)

An axis-shifting stereo camera, which is characterized by comprising; two lens groups, an axis-shifting device and an image processor; the said axis-shifting stereo camera consists of two independent, identical, and parallel centerlines. The lens group is composed of one or two identical image sensors CCD or CMOS, and the two lens groups are respectively located along a straight line which is located on a plane formed by the center lines of the two lens groups and perpendicular to the center lines of the two lens groups. When the two lens groups are shifted, one or two image sensors remain stationary; for an axis-shifting stereo camera with one image sensor, two lenses The two images collected by the group are respectively imaged on the left half and right half of the imaging surface of an image sensor, and output an image in the shift left and right format; for a shift stereo camera with two independent image sensors, two The two images collected by each lens group are respectively imaged on the imaging surface of an image sensor in each lens group, and output two independent axis-shifting images; the said one type of axis-shifting device is to combine the two images of an axis-shifting stereo camera. One lens group is located on a plane formed by the center lines of the two lens groups and is perpendicular to the center lines of the two lens groups, respectively, and are translated in the direction opposite to each other by L=T÷(2A) distance Device; In the above formula, T is the distance between the eyes of a person, and A is the magnification of the screen. The axis-shifting stereo camera according to claim 1, wherein the one image processor is an image processor provided with one or two image processing chips ISP, a touch screen, a data memory, and also includes an integrated And a device composed of a same-screen chip loaded and executed by the processor that stores multiple instructions; Among them, a three-dimensional image same-screen instruction is based on the principle of equivalent convergence. When the screen magnification A and the three-dimensional depth Z of an object of interest in the real scene change according to the formula A=[T÷(F×t)]×Z, A stereo camera composed of two lens groups or cameras that are independent of each other, the same and the center lines are arranged parallel to each other. The convergence point of a stereo image of an object of interest in the real scene collected by the real scene is always kept on the screen; where F is the lens group or camera The focal length of the lens, t is the distance between the two lens groups or the center line of the camera lens; A kind of stereo image measurement instruction is to establish a focus on an object based on the geometric relationship between two independent lens groups or cameras and an object of interest in the real scene, the same and the center line is set parallel to each other. The relationship between the parallax of the left and right images of a point and the spatial coordinates of the point of interest in the real scene; establish a relationship between the area of the image on the surface of the object of interest and the actual area of the surface of the object of interest in the real scene; For an image in a tilt-shift left and right format, the spatial coordinates of a point of interest are; x=t×(XL+T/2)÷[T－(XR－XL)]－t/2 y=YL÷(m×A)=YR÷(m×A) z=(A×F×t)÷[T－(XR－XL)] For two independent shift images on the left and right, the spatial coordinates of a point of interest are; x=t×(XL+T/2)÷[T－(XR－XL)]－t/2 y=YL÷(m×A)=YR÷(m×A) z=(A×F×t)÷[T－(XR－XL)] Among them, XL, XR, YL, and YR are the horizontal and vertical coordinates of the left and right image screenshots of the left and right image screenshots in a left and right format image screenshot of the left and right images of a point of interest, or two independent image screenshots of the left and right. ; M is the magnification of the lens group or camera lens; A three-dimensional image positioning and tracking command is based on the principle of equivalent convergence, a focus point or a straight line of interest in the left and right images in the real scene collected by two independent, identical and centerline lens groups or cameras. After the position of the left or right image in a left and right format image screenshot or two separate left and right image screenshots in the left image screenshot or right image screenshot is located, locate and track the point of interest or focus on the straight right image or The position of the left image in the right image screenshot or the left image screenshot in the same left and right format image screenshot or two separate left and right image screenshots; An equivalent convergence point reset command is to set the object as a new focus object through the stereo image of an object on the screen during the playback of the stereo image, and then use the stereo image of the new focus object to converge the equivalent of the stereo camera The point is reset to the new object of interest.

Images