JP2005351645A - Surface-damage detecting method, detector therefor and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a surface-damage detecting method for effectively extracting the surface damages of an object to be inspected. <P>SOLUTION: A blurred image forming means 10 is constituted so as to form a plurality of blurred images Sm (m=1-n, n≥2) that differ in the frequency response characteristics, on the basis of the original image S0, and a band-restricting image forming means 20 is constituted so as to perform subtraction of images of adjacent frequency bands, using the original image S0 and the respective blurred images Sm to form (n) band restricting images Tm. A damage component extracting means 30 applies nonlinear conversion to the respective band restricting images Tm, to obtain a plurality of converted images and to respectively add the pixel values of the pixels, corresponding to those converted images, to obtain the damage component Q of the original image S0. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a method and apparatus for detecting a surface flaw on a surface of an inspection object, specifically a surface flaw detection method and apparatus for detecting a flaw on the surface of an inspection object using an object image obtained by imaging the inspection object. It is about the program.

  Detection of a surface flaw on an inspection object is required in various fields. For example, detection of small scratches on the surface of metal cutting products, detection of fine paint unevenness on the surface of automobiles, detection of damage to agricultural products such as citrus fruits, and the like are performed. In the field of photographic images, it is not small scratches or so-called scratches on the face or the like, but it is also possible to detect and repair spots and wrinkle components on the face or the like. Visual inspection by human vision is easy to produce individual differences, and even for the same inspector, it can be obtained by imaging the inspection object because the inspection results often vary depending on the degree of eye fatigue. A method has been proposed in which image processing is performed on a target object image to automatically detect a surface flaw on the inspection target object.

  Patent Document 1 uses the fact that when damaged parts of epidermis oil vesicles such as citrus are irradiated with ultraviolet light, the damaged parts are excited to emit light, and the surface of agricultural products such as citrus is irradiated with ultraviolet light to damage the oil vesicles. A method has been proposed in which a damaged part is detected by performing image processing on an object image obtained by exciting and emitting the part and capturing the light with a camera or a light detection sensor through a filter.

  Further, since a fine flaw on the surface of the inspection object appears as a signal with a small gray level in the high-frequency component in the object image, it can be considered that the noise removal technique is applied to the detection of the surface flaw. For example, a low-pass filter (LPF) that is conventionally used for noise removal is applied to an object image to obtain a fine flaw component or a suppressed image with suppressed noise, and a flaw is obtained by subtracting the suppressed image from the original object image. Ingredients can be obtained. However, the LPF can suppress the scratch component from the image, but also blurs the edge portion. Therefore, the edge component remains in the scratch component obtained by subtracting the suppression image from the original object. Therefore, there is a problem that the detection accuracy is not good.

A noise removal technique that removes the noise component and does not degrade the edge component includes a method using an ε-filter (see Non-Patent Document 1). The ε-filter basically operates so as to subtract a value obtained by applying a nonlinear function to the amount of change in amplitude level in the high frequency component from the original image signal. This nonlinear function is a function that sets the output to 0 when the amplitude of the signal is larger than a predetermined threshold (here, ε ₀ ). That is, when the ε-filter is applied, the output of the non-linear function is 0 in a part of the image where the amplitude is larger than the threshold value, and the original signal of the part is retained in the processed image. In a region where the amplitude is equal to or smaller than the threshold value, in the processed image, the signal value of the region is a value obtained by subtracting the output of the nonlinear function (its absolute value is greater than 0) from the original signal value. When this ε-filter is applied to the detection of surface flaws, it is possible to prevent edge components from remaining as flaw components, and to increase the accuracy of flaw component detection.
Japanese Patent Laid-Open No. 06-129987 Harashima et al., “Ε-Separation Nonlinear Digital Filter and its Applications”, IEICE 1982, J65-A, No. 4, pp 297-304

  However, the method described in Patent Document 1 requires a device that irradiates an inspection object with ultraviolet light and a special imaging device (such as a camera or an optical sensor) that captures excitation light, and ultraviolet light such as a human face. It cannot be applied to an inspection object for which it is desired to avoid irradiation.

  Moreover, although many fine scratch components on the surface of the inspection object exist in the high frequency band, they exist over each frequency band from the high frequency band to the low frequency band. The above-described detection method using the ε-filter has a problem in that the flaw component cannot be completely detected because the flaw component is extracted only in one frequency band.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide a method and apparatus capable of effectively detecting a scratch component on the surface of an inspection object without requiring a special apparatus, and a program therefor. It is.

The surface flaw detection method of the present invention creates a plurality of band limited images representing components for each of a plurality of frequency bands of the object image based on the object image obtained by imaging the inspection object,
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is equal to or smaller than the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Apply to each pixel value of the image to get multiple converted images,
An added pixel value obtained by adding pixel values of corresponding pixels of the plurality of converted images is acquired as a pixel value of a flaw component image pixel representing a flaw component on the surface of the inspection object. Is.

  Here, the “object image” means not only a digital image obtained by imaging a subject with a digital camera, but also an image on a silver halide photographic film or a printed material (for example, a print) by a reading device such as a scanner. The obtained digital image is also included. Hereinafter, for convenience of explanation, an image simply means an object image in the present invention.

  In addition, “acquiring an added pixel value obtained by adding pixel values of corresponding pixels of a plurality of converted images as a pixel value of a flaw component image on the surface of the inspection object” as a result Means to obtain the added pixel value as a pixel value of a pixel of the flaw component image, and the added pixel value is obtained by directly adding pixel values of corresponding pixels of the plurality of converted images. Alternatively, for example, in a system having a step of obtaining an image (repaired image) obtained by removing the scratch component from the object image, the plurality of converted images are obtained from pixel values of pixels of the object image. Of course, the pixel value of the pixel of the repaired image obtained by subtracting the pixel value of the corresponding pixel may be obtained by subtracting the pixel value of the pixel of the object image.

  In the surface flaw detection method of the present invention, a value obtained by multiplying the addition pixel value corresponding to the pixel by a coefficient determined according to the pixel value of the pixel of the object image is used as the value of the flaw component image. It is preferable to obtain the pixel value of the pixel, and a specific method of determining the coefficient may be a method based on the type of the object to be detected, the degree of detection of the scratch component, and the like.

  For example, since a flaw component in an image usually appears as a dark portion in an image (that is, a portion with a small pixel value such as a luminance value), the smaller the pixel value of the object image, the smaller the image value (in contrast, the image value). If a coefficient is used as it is smaller, the edge component can be prevented from being detected as a flaw component, and more flaw components can be detected.

  Also, in the present invention, components such as wrinkles and spots in the normal sense are also regarded as scratch components on the surface of the skin such as the face, and the surface scratch detection method of the present invention is applied to detection of wrinkles and stain components in facial photographic images and the like. can do. In a photographic image, wrinkles and stain components are present in lighter skin parts such as hair, so that the larger the pixel value, for example, the luminance value (or the smaller the pixel value, the smaller) is used as the coefficient. Then, it is possible to detect wrinkles and stain components, and to prevent erroneous detection of hair and the like as wrinkles and stain components.

  The nonlinear conversion process is preferably a process in which the absolute value of the output value is substantially constant with respect to an input value whose absolute value is greater than the predetermined threshold. Here, an input value larger than the predetermined threshold value can be considered as a signal value of a part that is not a flaw component, and if the nonlinear transformation process is performed so that the output value is substantially constant, the calculation is performed. Time can be saved and scratch components can be extracted efficiently. The absolute value of the output value of the input value whose absolute value is larger than the predetermined threshold value is a value within the range from zero to the maximum output value (output value corresponding to the threshold value), and the larger the value, the more flaw components are extracted. The smaller the value, the more the edge component can be prevented from remaining.

  Further, the predetermined threshold has a property of a threshold for distinguishing a flaw component from an edge component, and the same value may be used for each of the band limited images. It is preferable that it is determined according to the frequency band.

  The pixel value of the pixel of the flaw component image obtained by the surface flaw detection method of the present invention represents the flaw component image. In addition to the flaw component, the flaw component image is caused by an image sensor such as a CCD. Since noise and the like are included, it is desirable to further apply a smoothing filter such as a median filter to the flaw component image as the surface flaw component detection method of the present invention. The obtained flaw component image is obtained by removing noise components other than the flaw component.

The surface flaw detection apparatus of the present invention creates a band limited image that creates a plurality of band limited images representing components for a plurality of frequency bands of an object image based on the object image obtained by imaging the inspection object. Creating means;
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is equal to or smaller than the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Non-linear conversion means for applying to each pixel value of the image to obtain a plurality of converted images;
Flaw component acquisition means for acquiring an added pixel value obtained by adding pixel values of corresponding pixels of the plurality of converted images as a pixel value of a flaw component image pixel representing a flaw component on the surface of the inspection object; It is characterized by having.

  The flaw component acquisition means obtains a pixel value of a pixel of the flaw component image by multiplying the addition pixel value corresponding to the pixel by a coefficient determined according to a pixel value of the pixel of the object image. It is preferable to acquire as.

  The program of the present invention causes a computer to execute the surface flaw detection method of the present invention.

  The surface flaw detection method and apparatus of the present invention and the program therefor perform non-linear transformation processing on a plurality of band limited images representing components of a plurality of different frequency bands of the image created based on the object image. To obtain a plurality of converted images. This non-linear conversion process is a process for setting the absolute value of the output value to be equal to or smaller than the absolute value of the input value. While the absolute value of the value is increased, for an input value whose absolute value is greater than a predetermined threshold, the absolute value of the output value is equal to or less than the absolute value of the output value corresponding to the predetermined threshold. The converted image obtained by this nonlinear conversion processing represents a component having a small amplitude such as a flaw or noise in the frequency band corresponding to the converted image. In the present invention, since the added pixel value obtained by adding the pixel values of the corresponding pixels of the plurality of converted images is acquired as the pixel value of the flaw component image of the inspection object, Scratch components in a plurality of frequency bands can be detected effectively.

  Further, it can be applied to all types of inspection objects without requiring ultraviolet irradiation or a special imaging device.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of an image processing apparatus according to an embodiment of the present invention. Note that the image processing apparatus according to the present embodiment performs a process of extracting a flaw component from an input inspection object image, and executes a processing program read into an auxiliary storage device on a computer (for example, a personal computer). It is realized by doing. Further, this processing program is stored in an information storage medium such as a CD-ROM or distributed via a network such as the Internet and installed in a computer.

  Further, since the image data represents an image, the following description will be given without particularly distinguishing the image from the image data.

  As shown in FIG. 1, the image processing apparatus according to the present embodiment creates a plurality of blurred images S1, S2,... Sn (n: integer greater than or equal to 2) with different frequency response characteristics of the original image S0. Image creating means 10, band limited image creating means 20 for creating a plurality of band limited images T1, T2,... Tn using original image S0 and blurred images S1, S2,. The wound component extraction means 30 for extracting the wound component Q using T1, T2,... Tn, and the wound repair means for removing the wound component Q from the original image S0 to obtain the wound repaired image S′0 (Y1) 40, a smoothing means 50 for performing a smoothing process on the scratch component Q to obtain a scratch component Q ′ from which noise has been removed, and an output means 60 for outputting the scratch component Q ′. The processing is performed in the luminance space. The processing device performs YCC conversion on the input image D0 (R0, G0, B0) to obtain the luminance Y0 of the image D0 (the original image S0 is composed of the luminance component Y0), the color difference Cb0, The YCC conversion means 1 for obtaining Cr0, the pixel value Y1 of the wound repaired image S′0 obtained by the wound repair means 40, and the color differences Cb0 and Cr0 obtained by the YCC conversion means 1 are combined and processed. A synthesizing unit 44 for obtaining D1 (Y1, Cb0, Cr0) and a display unit 48 for displaying the processed image D1 are also provided. Hereinafter, details of each component of the present embodiment will be described.

  The YCC conversion unit 1 converts the R, G, and B values of the image data D0 into the luminance value Y and the color difference values Cb and Cr according to the following equation (1).

Y = 0.2990 × R + 0.5870 × G + 0.1140 × B
Cb = −0.1687 × R−0.3313 × G + 0.5000 × B + 128 (1)
Cr = 0.5000 × R−0.4187 × G−0.0813 × B + 128

The blur image creating unit 10 creates a plurality of blur images using the luminance value Y0 obtained by the YCC conversion unit 1. FIG. 2 is a block diagram illustrating a configuration of the blurred image creation unit 10. As shown in the figure, the blurred image creation means 10 includes a filtering means 12 that performs filtering processing to obtain filtered images B1, B2,... Bn, an interpolation means 14 that performs interpolation processing on each filtering image, and filtering. The control means 16 which controls the means 12 and the interpolation means 14 is provided. The filtering means 12 performs a filtering process using a low-pass filter. As the low-pass filter, for example, a filter F substantially corresponding to a 5 × 1 grid-shaped one-dimensional Gaussian distribution as shown in FIG. 3 is used. be able to. This filter F has σ = 1 in the following equation (2).

The filtering unit 12 performs a filtering process on the entire image to be processed by performing a filtering process on the image to be processed in the x direction and the y direction of the pixel using such a filter F.

FIG. 4 shows details of processing that the control unit 16 of the blurred image creation unit 10 causes the filtering unit 12 and the interpolation unit 14 to perform. As shown in the figure, the filtering unit 12 first performs a filtering process on the original image S0 (Y0) every other pixel using the filter F shown in FIG. A filtering image B1 (Y1) is obtained by this filtering process. The size of the filtering image B1 is 1/4 of the size of the original image S0 (1/2 in the x direction and y direction, respectively). Next, the filtering unit 12 performs the filtering process by the filter F every other pixel on the filtering image B1 (Y1) to obtain the filtering image B2 (Y2). The filtering means 12 repeats the filtering process by such a filter F, and obtains n filtered images Bk (k = 1 to n). The size of the filtering image Bk has become a ^{1/2 2K} of the size of the original image S0. FIG. 5 shows frequency characteristics of each filtered image Bk obtained by the filtering unit 12 when n = 3 as an example. As shown in the figure, the response of the filtered image Bk is such that the higher the k, the higher the frequency component is removed.

  In the image processing apparatus according to the present embodiment, the filtering unit 12 performs the filtering process on the x direction and the y direction of the image by the filter F shown in FIG. A filtering process may be performed on the original image S0 and the filtered image Bk at a time using a × 5 two-dimensional filter.

The interpolation unit 14 performs an interpolation process on each filtering image Bk obtained by the filtering unit 12 so that the size of each filtering image Bk is the same as that of the original image S0. There are various interpolation processing methods such as a B-spline method. In this embodiment, the filtering unit 12 uses a filter F based on a Gaussian signal as a low-pass filter. A Gaussian signal is also used as an interpolation coefficient for performing the interpolation calculation, and this interpolation coefficient is approximated to σ = 2 ^K−1 in the following equation (3).

When interpolating the filtered image B1, since k = 1, σ = 1. The filter for performing interpolation when σ = 1 in the above equation (3) is a 5 × 1 one-dimensional filter F1 as shown in FIG. The interpolating means 14 first enlarges the filtered image B1 to the same size as the original image S0 by interpolating pixels with a value of 0 every other pixel one by one with respect to the filtered image B1, and then for the enlarged image Then, a filtering process is performed by the filter F1 shown in FIG. 7 to obtain a blurred image S1. The blurred image S1 has the same number of pixels as the original image S0, that is, the same size as the original image S0.

  Here, the filter F1 shown in FIG. 7 is a 5 × 1 filter. However, before applying the filter F1, pixels having a value of 0 are interpolated with respect to the filtered image B1 every other pixel. The interpolation processing by the means 14 is substantially performed by two types of filters: a 2 × 1 filter (0.5, 0.5) and a 3 × 1 filter (0.1, 0.8, 0.1). Equivalent to filtering.

  When the interpolation means 14 performs interpolation on the filtered image B2, k = 2, so σ = 2. In the above equation (3), the filter corresponding to σ = 2 is an 11 × 1 one-dimensional filter F2 shown in FIG. The interpolation unit 14 first expands the filtering image B2 to the same size as the original image S0 by interpolating three pixels each having a value of 0 every other pixel with respect to the filtering image B2, and then expanding the image Is subjected to a filtering process by a filter F2 shown in FIG. 8 to obtain a blurred image S2. The blurred image S2 has the same number of pixels as the original image S0, that is, the same size as the original image S0.

  Similarly, the filter F2 shown in FIG. 8 is an 11 × 1 filter, but before applying the filter F2, three pixels each having a value of 0 are interpolated with respect to the filtered image B2 by three. Therefore, the interpolation processing by the interpolation means 14 is substantially performed by a 2 × 1 filter (0.5, 0.5) and a 3 × 1 filter (0.3, 0.65, 0.05), ( This is equivalent to filter processing using four types of filters (0.3, 0.74, 0.13) and (0.05, 0.65, 0.3).

In this way, the interpolation means 14 interpolates (2 ^K −1) pixels each having a value of 0 every other pixel with respect to each filtering image Bk, so that the filtering image Bk is the same as the original image S0. For a filtered image Bk that is enlarged to size and interpolated with pixels having a value of 0, the length created based on the above equation (3) is (3 × 2 ^K −1). A blurred image Sk is obtained by performing a filtering process.

  FIG. 9 shows the frequency characteristics of each blurred image Sk obtained by the blurred image creating means 10 when n = 3 as an example. As shown in the figure, the blurred image Sk is such that the higher the k, the higher frequency components of the original image S0 are removed.

  The band-limited image creating means 20 uses the blurred images S1, S2,... Sn obtained by the blurred image creating means 10 and components for each of a plurality of frequency bands of the original image S0 according to the following equation (4). .., Tn are generated.

Tm = S (m−1) −Sm (4)
Where m is an integer between 1 and n

FIG. 10 shows the frequency characteristics of each band limited image Tm obtained by the band limited image creating means 20 when n = 3 as an example. As illustrated, the band limited image Tm represents a component in the low frequency region of the original image S0 as m increases.

  The wound component extraction means 30 extracts the wound component Q using each band limited image Tm (m = 1 to n) obtained by the band limited image creation means 20, and as shown in FIG. Non-linear transformation is performed on each band limited image Tm (m = 1 to n), and a converted image of each band limited image Tm, that is, a flaw or noise component (hereinafter referred to as total) in a frequency band corresponding to each band limited image Tm. .., Qn and the flaw components Q1, Q2,... Qn in each frequency band are added to obtain the flaw component Q in the original image S0. Is. The non-linear conversion performed by the non-linear conversion means 32 is a process of setting the output value to be equal to or lower than the input value, and for an input value equal to or lower than a predetermined threshold value, the larger the input value, the larger the output value. For an input value larger than the predetermined threshold, the output value is equal to or lower than the output value corresponding to the predetermined threshold. In this embodiment, the process is performed by a function f as shown in FIG. Is. In the figure, the broken line indicates a function in which the output value = the input value, that is, the slope is 1. As illustrated, the nonlinear transformation function f used for the nonlinear transformation means 32 in the flaw component extraction means 30 of the present embodiment has a slope of 1 when the absolute value of the input value is smaller than the first threshold Th1, When the absolute value of the input value is greater than or equal to the first threshold value and less than or equal to the second threshold value Th2, the slope is smaller than 1. When the absolute value of the input value is greater than the second threshold value Th2, the output value is its absolute value. It is a function whose value is a constant value M smaller than the absolute value of the input value. The function f may be the same for each band limited image, but may be different for each band limited image.

  The nonlinear conversion means 32 uses the luminance value of each band limited image as an input value, performs nonlinear conversion on each band limited image using the nonlinear conversion function f shown in FIG. The scratch component Qm (m = 1 to n) in the frequency band to which each band limited image corresponds is acquired.

The adding means 34 adds the flaw component Qm (m = 1 to n) in each frequency band obtained by the nonlinear conversion means 32 according to the following equation (5) to obtain the flaw component Q of the original image S0.

Here, the coefficient β is β (S0), that is, determined according to the luminance value Y0 of the pixel of the original image S0. This coefficient β is determined according to the pixel value, here the luminance value Y0, based on the degree to detect the flaw component, the type of detection object, and the like. For example, a flaw component can be effectively detected by using a coefficient β having a smaller value for a pixel having a larger luminance value Y0. Further, when the inspection target is a human face, it is possible to prevent a hair part from being detected as a scratch component by using a coefficient β having a larger value for a pixel having a larger luminance value.

  The smoothing unit 50 applies a smoothing filter such as an integration filter or a media filter to the scratch component Q obtained by the scratch component extraction unit 30 to smooth the scratch component from which noise has been removed. Get Q '.

  The output means 60 outputs the scratch component Q ′ (including output as data and display as an image). Here, the output means 60 may output the scratch component Q ′ as it is, or the pixel value (in this case, the luminance value) of the scratch component Q ′, for example, based on the use of the extracted scratch component. The output may be performed by performing processing such as obtaining and outputting the position of a pixel other than 0 as the position of a flaw.

  The wound repair means 40 subtracts the wound component Q extracted by the wound component extraction means 30 from the original image S0 (Y0) to obtain a wound repaired image S′0 (Y1).

  The synthesizing unit 44 synthesizes the luminance value Y1 obtained by the flaw repairing unit 40 and the color difference values Cb0 and Cr0 of the image D0 obtained by the YCC conversion unit 1 to process the processed image D1 (Y1, Cr0, Cb0). Get.

  The display means 48 displays the processed image D1 obtained by removing the scratch component and allows the operator to confirm the display. For example, a monitor or the like can be used.

  As described above, the image processing apparatus according to the present embodiment creates a plurality of band limited images Tm (m = 1 to n, n ≧ 2) representing components for a plurality of different frequency bands of the original image S0 (Y0). A non-linear conversion process is performed on these band limited images to obtain a plurality of converted images (scratch components). Then, by multiplying the pixel values of the plurality of converted images by a coefficient β determined according to the luminance value of the original image S0 and adding them, the flaw component of the original image S0 is obtained. Scratch components in a plurality of frequency bands can be extracted effectively.

  Further, the image processing apparatus of the present embodiment, for example, thresholds Th1 and Th2 for determining the nonlinear transformation function f shown in FIG. 12, the slope of the curve of the nonlinear transformation function f, the magnitude of the output value M, and equations (5) to (5) Conveniently, it is possible to change the degree to detect the scratch component by changing only the coefficient β in the equation (9), and to deal with different types of inspection objects.

 The preferred embodiments of the present invention have been described above. However, the surface flaw detection method and apparatus of the present invention and the program therefor are not limited to the above-described embodiments, and various methods are possible without departing from the gist of the present invention. Increase, decrease, change can be added.

  For example, in the image processing apparatus of the present embodiment, the band limited image creating unit 20 uses the original image S0 and the blurred image Sk (k = 1 to n, n ≧ 2) to generate the band limited image according to the above equation (4). The processing performed in the band limited image creating means 20 and the flaw component extracting means 30 can be expressed by the following formula (6). For example, the following formula (7) or formula (8) ) Or equation (9), the band-limited image creation means 20 and the flaw component extraction means 30 may be processed. That is, as in the processing in the image processing apparatus of the present embodiment represented by Expression (6), using the original image and each blurred image, the images in the frequency bands adjacent to each other (the original image S0 is the frequency of the blurred image S1 and the frequency The band-limited image may be obtained by subtracting the band-limited images), but as shown in Expression (7), the band-limited image is obtained by subtracting all the blurred images and the original image. Alternatively, as shown in Expression (8), the band-limited image may be obtained by subtracting the blurred images in the frequency bands adjacent to each other without using the original image, or Expression (9). ), The band-limited image is obtained by subtraction of the blurred image S1 and all the blurred images Sm (m = 2 to n, n ≧ 3) excluding the blurred image S1 without using the original image. May be.

Also, the method for creating the band limited image is to create a blurred image of the original image, and to create the blurred image of the original image and / or the blurred image as in the methods represented by the above-described equations (4) and (6) to (9). The method is not limited to the method of creating using an image, and any method may be used as long as an image representing a plurality of different frequency band components of the original image can be created.

  In the image processing apparatus of this embodiment, the scratch component Qm in each frequency band is directly added to obtain the scratch component Q, and the scratch component Q is subtracted from the original image S0 (Y0) to repair the scratched image S ′. 0 (Y1) is obtained, but after the wound component Qm in each frequency band is subtracted from the original image S0 one by one to obtain the wound repaired image S′0, the wound repaired image S ′ is obtained from the original image S0. The scratch component Q may be obtained by subtracting 0.

  Further, in the image processing apparatus of the present embodiment, the operator is confirmed by displaying the result of repairing the image based on the extracted wound component Q, but the purpose is only the extraction of the wound component. In some cases, it is not necessary to perform the process of repairing the image.

  Further, the image processing apparatus according to the present embodiment extracts the scratch component in the luminance space. However, for example, it may be performed for all of the R, G, and B components.

1 is a block diagram showing a configuration of an image processing apparatus according to an embodiment of the present invention. 1 is a block diagram showing the configuration of a blurred image creation unit 10 in the image processing apparatus shown in FIG. The figure which shows the example of the one-dimensional filter F which the filtering means 12 in the blur image creation means 10 shown in FIG. 2 uses The figure which shows the process performed in the blurred image preparation means 10 shown in FIG. The figure which shows the frequency characteristic of the filtering image Bk produced by the filtering means 12 in the blur image creation means 10 shown in FIG. The figure which shows the example of the two-dimensional filter which the filtering means 12 in the blur image creation means 10 shown in FIG. 2 uses The figure which shows the example of the filter F1 which the interpolation means 14 in the blur image creation means 10 shown in FIG. 2 uses for interpolation of the filtering image B1 The figure which shows the example of the filter F2 which the interpolation means 14 in the blur image creation means 10 shown in FIG. 2 uses for interpolation of filtering image B2 The figure which shows the frequency characteristic of the blur image Sk created by the blur image creation means 10 shown in FIG. The figure which shows the frequency characteristic of the band limited image Tk produced by the band limited image preparation means 20 in the image processing apparatus of embodiment of this invention shown in FIG. The block diagram which shows the structure of the flaw component extraction means 30 in the image processing apparatus of embodiment of this invention shown in FIG. The figure which shows the example of the nonlinear function f which the nonlinear transformation means 32 in the flaw component extraction means 30 shown in FIG. 11 uses

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 YCC conversion means 10 Blurred image creation means 12 Filtering means 14 Interpolation means 16 Control means 20 Band-limited image creation means 30 Wound component extraction means 32 Non-linear conversion means 34 Addition means 40 Wound repair means 44 Composition means 48 Display means 50 Smoothing Means 60 Output means Y Luminance value Cb, Cr Color difference value

Claims (5)

Based on the object image obtained by imaging the inspection object, create a plurality of band limited images representing components for each of a plurality of frequency bands of the object image,
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is equal to or smaller than the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Apply to each pixel value of the image to get multiple converted images,
An added pixel value obtained by adding pixel values of corresponding pixels of the plurality of converted images is acquired as a pixel value of a flaw component image pixel representing a flaw component on the surface of the inspection object. Surface scratch detection method.   A value obtained by multiplying the addition pixel value corresponding to the pixel by a coefficient determined according to a pixel value of the pixel of the object image is acquired as a pixel value of the pixel of the flaw component image. The surface flaw detection method according to claim 1. Band-limited image creating means for creating a plurality of band-limited images representing components for each of a plurality of frequency bands of the object image based on the object image obtained by imaging the inspection object;
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is equal to or smaller than the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Non-linear conversion means for applying to each pixel value of the image to obtain a plurality of converted images;
Flaw component acquisition means for acquiring an added pixel value obtained by adding pixel values of corresponding pixels of the plurality of converted images as a pixel value of a flaw component image pixel representing a flaw component on the surface of the inspection object; A surface flaw detection device comprising:   The flaw component acquisition means multiplies the addition pixel value corresponding to the pixel by a coefficient determined according to the pixel value of the pixel of the object image, and obtains a pixel value of the pixel of the flaw component image The surface flaw detection device according to claim 3, wherein the surface flaw detection device is acquired as: Based on the object image obtained by imaging the inspection object, a process of creating a plurality of band limited images representing components for each of a plurality of frequency bands of the object image;
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is less than or equal to the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Processing to obtain a plurality of converted images by applying to each pixel value of the image;
A process of acquiring an added pixel value obtained by adding pixel values of corresponding pixels of the plurality of converted images as a pixel value of a flaw component image pixel representing a flaw component on the surface of the inspection object; A program characterized by being executed.

Images