TWI841130B - Method for detecting lane line and related devices - Google Patents

Abstract

The present application provides a method for detecting a lane line and related devices. The method includes: converting an obtained foreground image into an aerial view, moving a horizontal sliding window on the aerial view, fitting a main lane line utilizing non-zero pixel points within the horizontal sliding window; in response that there are confidence coefficients of a continuous preset number of horizontal sliding windows are less than a preset threshold, determining a previous horizontal sliding window before the continuous preset number of horizontal sliding windows as a prior horizontal sliding window, and determining an end of the main lane line according to the prior horizontal sliding window; moving a vertical sliding window on the aerial view, and generating a target curve by fitting non-zero pixel points within the vertical sliding window; in response that the end of the main lane line is on the target curve, determining that the target curve is a stop line. This application can improve the accuracy of lane line recognition.

Description

Lane line detection method and related equipment

The present invention relates to the field of vehicle technology, and in particular to a lane line detection method and related equipment.

Lane line detection is an important technology in unmanned driving or assisted driving scenarios. Lane line detection refers to the detection of traffic signs (i.e. lane lines) on the road. Lane line detection can be used to determine whether the vehicle has deviated during driving. If the lane line cannot be accurately identified, it will affect the safe driving of the vehicle. Therefore, in smart driving, lane lines need to be accurately identified.

In view of the above, it is necessary to provide a lane line detection method and related equipment that can solve the technical problem of being unable to identify the stop line.

The present application provides a lane line detection method, the method comprising: obtaining a foreground image of a vehicle; converting the foreground image into a bird's-eye view, and establishing a horizontal bar graph corresponding to the bird's-eye view; using the peak of the horizontal bar graph as the starting point of a moving horizontal sliding window, and generating a main lane line by fitting non-zero pixel points in the horizontal sliding window; calculating the confidence of each horizontal sliding window; if there are a continuous preset number of horizontal sliding windows whose confidence is less than a preset threshold value, determining the lane line detection method; The previous horizontal sliding window of the preset number of the horizontal sliding windows is used as the previous horizontal sliding window, and the end point of the lane line of the main lane line is determined according to the previous horizontal sliding window; a vertical bar graph corresponding to the bird's-eye view is established, and the peak of the vertical bar graph is used as the starting point of the moving vertical sliding window; a target curve is generated by fitting the non-zero primitive points in the vertical sliding window, and if the end point of the lane line is located on the target curve, the target curve is determined to be a stop line.

In some optional embodiments, the foreground image is converted into a bird's-eye view, including: performing distortion correction on the foreground image to obtain a corrected image; taking each non-zero pixel point in the corrected image as a target point, calculating the coordinates of the target point using a coordinate conversion formula to obtain an inverse perspective transformation matrix; and converting the corrected image into the bird's-eye view using the inverse perspective transformation matrix.

In some optional embodiments, the method of using the peak of the horizontal bar graph as the starting point for moving the horizontal sliding window and generating the main lane line by fitting the non-zero pixel points in the horizontal sliding window includes: using the peak of the horizontal bar graph as the starting point for moving the horizontal sliding window, calculating the horizontal coordinates of all non-zero pixels in the previous horizontal sliding window of the current horizontal sliding window, and calculating the horizontal coordinates of all non-zero pixels in the previous horizontal sliding window of the current horizontal sliding window. The average value is used as the first horizontal coordinate average value; the horizontal window center of the current horizontal sliding window is determined according to the first horizontal coordinate average value, and the position of the current horizontal sliding window is determined according to the horizontal window center; the non-zero primitive points in the current horizontal sliding window and all horizontal sliding windows before the current horizontal sliding window are fitted to generate the main lane line.

In some optional embodiments, the calculating the confidence of each of the horizontal sliding windows includes: inputting the bird's-eye view containing the horizontal sliding window into a preset deep learning neural network model, calculating the similarity between the primitive features and the sample features of each primitive in each of the horizontal sliding windows; determining the confidence of each of the horizontal sliding windows according to the similarity, wherein the similarity is proportional to the confidence.

In some optional embodiments, the method of using the crest of the vertical bar graph as the starting point for moving the vertical sliding window and generating a target curve by fitting non-zero pixel points in the vertical sliding window includes: using the crest of the vertical sliding window as the starting point for moving the vertical sliding window, calculating the horizontal coordinates of all non-zero pixels in the previous vertical sliding window of the current vertical sliding window, and calculating the horizontal coordinates of all non-zero pixels in the previous vertical sliding window of the current vertical sliding window. The average value is used as the second horizontal coordinate average value; the vertical window center of the current vertical sliding window is determined according to the second horizontal coordinate average value, and the position of the current vertical sliding window is determined according to the vertical window center; the non-zero primitive points in the current vertical sliding window and all vertical sliding windows before the current vertical sliding window are fitted to generate the target curve.

In some optional embodiments, if the end point of the lane line is located on the target curve, determining that the target curve is a stop line includes: calculating the matching degree between a plurality of longitudinal sliding windows corresponding to the target curve and the previous transverse sliding window; if any matching degree exceeds a preset matching degree, determining that the end point of the lane line is located on the target curve; and determining the position of the stop line based on the position of the target curve and the position of the end point of the lane line.

In some optional embodiments, after determining that the target curve is a stop line, the method further includes: filtering the continuous preset number of horizontal sliding windows.

In some optional embodiments, the method further includes: if the end point of the lane line does not exist on the target curve, determining that the stop line does not exist in the foreground image, and outputting a warning to indicate that the main lane line will not be recognized.

This application also provides a vehicle-mounted device, which includes a processor and a memory, and the processor is used to implement the lane line detection method when executing a computer program stored in the memory.

This application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the lane line detection method is implemented.

In the lane line detection method provided in the present application, in order to identify the main lane line and the stop line, first, the foreground image of the vehicle is obtained, the foreground image is converted into a bird's-eye view, a horizontal bar graph corresponding to the bird's-eye view is established, the peak of the horizontal bar graph is used as the starting point of the moving horizontal sliding window, and the main lane line is generated by fitting the non-zero pixel points in the horizontal sliding window, thereby improving the accuracy of identifying the main lane line. Then, the confidence of each horizontal sliding window is calculated, and the confidence of each horizontal sliding window is compared with the size of the preset valve value. If there are a continuous preset number of horizontal sliding windows whose confidence is less than the preset valve value, the previous horizontal sliding window of the continuous preset number of horizontal sliding windows is used as the previous horizontal sliding window, and the end point of the lane line is determined according to the previous horizontal sliding window. The confidence is used to determine the end point of the lane line, which further improves the accuracy of identifying the main lane line. Finally, a vertical bar graph corresponding to the bird's-eye view is established, and the peak of the vertical bar graph is used as the starting point of the moving vertical sliding window. The target curve is generated by fitting the non-zero pixel points in the vertical sliding window. If the end point of the lane line is located on the target curve, the target curve is determined to be the stop line. This application can improve the accuracy of identifying lane lines (including main lane lines and stop lines) in the acquired images by establishing horizontal and vertical bar graphs, thereby further improving driving safety.

1: Vehicle-mounted device

10: Communication bus

11: Storage

12: Processor

13: Shooting equipment

S21~S26: Steps

Figure 1 is a schematic diagram of a vehicle-mounted device provided in an embodiment of the present application.

Figure 2 is a flow chart of the lane line detection method provided in an embodiment of this application.

Figure 3 is a schematic diagram of the confidence provided by an embodiment of this application.

Figure 4 is a schematic diagram of the confidence provided by an embodiment of this application.

Figure 5 is a flow chart of the lane line detection method provided in an embodiment of this application.

Figure 6 is a schematic diagram of lane lines provided in an embodiment of this application.

For ease of understanding, some explanations of concepts related to the embodiments of this application are given as examples for reference.

It should be noted that in this application, "at least one" means one or more, and "plurality" means two or more than two. "And/or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and/or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The terms "first", "second", "third", "fourth", etc. (if any) in the specification, claim and drawings of this application are used to distinguish similar objects, not to describe a specific order or precedence.

In order to better understand the lane line detection method and related equipment provided by the embodiment of this application, the application scenario of the lane line detection method of this application is first described below.

FIG1 is a schematic diagram of a vehicle-mounted device provided in an embodiment of the present application. The lane line detection method provided in the embodiment of the present application is applied to the vehicle-mounted device 1, and the vehicle-mounted device 1 includes, but is not limited to, a storage device 11, at least one processor 12, and a camera 13 connected to each other via a communication bus 10, and the camera 13 can be a vehicle-mounted camera device or an external vehicle camera device, such as a camera or a dash cam, to capture multiple images or videos in front of the vehicle.

Figure 1 is only an example of the vehicle-mounted device 1 and does not constitute a limitation on the vehicle-mounted device 1. The vehicle-mounted device 1 in actual application may include more or fewer components than shown in the figure, or combine certain components, or replace different components. For example, the vehicle-mounted device 1 may also include input and output devices, network access devices, etc.

In the embodiment of the present application, the vehicle-mounted device 1 is applied to a vehicle, for example, it can be a vehicle-mounted device in a vehicle (for example, a vehicle computer), or it can be an independent electronic device (for example, a computer, a mobile phone, a tablet computer, etc.) and can communicate and interact with the vehicle-mounted device to achieve control of the vehicle.

As shown in FIG2, it is a flow chart of a lane line detection method provided in an embodiment of the present application. The lane line detection method described in the present application is applied in a vehicle-mounted device (such as the vehicle-mounted device 1 in FIG1). According to different requirements, the order of the steps in the flow chart can be changed, and some steps can be omitted.

Step S21, obtaining the foreground image of the vehicle.

In an embodiment of the present application, the foreground image may be an image taken in the direction of the vehicle's travel. The foreground image in front of the vehicle may be photographed to obtain the foreground image. Alternatively, the scene in front of the vehicle may be photographed by video, and the foreground image may be obtained from the video.

When the vehicle-mounted device includes a camera, the first foreground image can be obtained by the camera of the vehicle-mounted device. When the vehicle-mounted device does not include a camera, the foreground image can be obtained by a camera on the vehicle (such as a dashcam).

Step S22, convert the foreground image into a bird's-eye view, and create a horizontal bar graph corresponding to the bird's-eye view.

In an embodiment of the present application, due to the angle, rotation, scaling, etc. of the shooting device during shooting, the first foreground image may be distorted (i.e., distorted), and the foreground image needs to be distorted.

The foreground image is distorted and a corrected image is obtained. Specifically, an image coordinate system is established for the foreground image, the first coordinate of each non-zero pixel point in the foreground image in the image coordinate system is obtained, the internal parameter of the shooting device is obtained, and the second coordinate corresponding to the first coordinate is determined according to the internal parameter and the first coordinate, wherein the second coordinate is a non-distorted coordinate. According to the first coordinate and the center coordinate of the foreground image, the distortion distance between the first coordinate and the center coordinate is determined. According to the gray value of each pixel point in the foreground image, the image complexity of the foreground image is calculated, and the correction parameter of the foreground image is determined according to the image complexity. According to the preset smoothing function, the smoothing coefficient corresponding to the distortion distance and the correction parameter is determined. According to the smoothing coefficient and the second coordinate, the first coordinate is smoothed and corrected to obtain a corrected image.

After obtaining the corrected image, each non-zero pixel point in the corrected image is taken as a target point, and the coordinates of the target point are calculated using a coordinate conversion formula to obtain an inverse perspective transformation matrix, and the corrected image is converted into a bird's-eye view using the inverse perspective transformation matrix. Specifically, the corrected image may be grayed, gradient valve value, color valve value and saturation valve value pre-processed, etc., to remove irrelevant lane line information in the corrected image to obtain a binary image, and the perspective effect is eliminated by using a perspective transformation by using the characteristic that lanes on the same road surface are approximately parallel, and each non-zero pixel point in the corrected image is taken as a target point, and the target point is calculated using a coordinate conversion formula to obtain an inverse transformation matrix, and the binary image is perspective transformed by using the inverse transformation matrix to obtain a bird's-eye view. Among them, the bird's-eye view is a three-dimensional picture drawn based on the perspective principle, using the high viewpoint method to look down at the undulations of the ground from a certain point at a high altitude. Compared with the plane map, it is more realistic.

After obtaining the bird's-eye view, a horizontal bar graph corresponding to the bird's-eye view is established, specifically including: establishing a horizontal bar graph for the non-zero pixel points corresponding to the lower half of the bird's-eye view, and obtaining the peaks of the horizontal bar graph according to the accumulated value of the number of non-zero pixel points in each column, for example, including the first peak and the second peak.

Step S23, taking the peak of the horizontal bar graph as the starting point for moving the horizontal sliding window, and generating the main lane line by fitting the non-zero pixel points in the horizontal sliding window.

In an embodiment of the present application, based on the first wave crest and the second wave crest obtained, a lane line is searched on a bird's-eye view using a horizontal sliding window, specifically including: using the first wave crest as the starting point for moving the horizontal sliding window, calculating the average value of the horizontal coordinates of all non-zero pixels in the previous horizontal sliding window of the current horizontal sliding window as the first horizontal coordinate average value, determining the horizontal window center of the current horizontal sliding window according to the first horizontal coordinate average value, determining the position of the current horizontal sliding window according to the horizontal window center, fitting the non-zero pixel points in the current sliding window and all horizontal sliding windows before the current horizontal sliding window, and generating the left main lane line.

Similarly, the horizontal sliding window is moved based on the second wave peak as the starting point, and the average value of the horizontal coordinates of all non-zero pixels in the previous horizontal sliding window of the current horizontal sliding window is calculated as the first horizontal coordinate average value. The horizontal window center of the current horizontal sliding window is determined according to the first horizontal coordinate average value, and the position of the current horizontal sliding window is determined according to the horizontal window center. The non-zero pixel points in the current sliding window and all horizontal sliding windows before the current horizontal sliding window are fitted to generate the right main lane line.

In the present application embodiment, in order to quickly simulate the main lane line, if the horizontal sliding window moves to an area with fewer non-zero pixel points within the range of the bird's-eye view, the horizontal sliding window is moved in a straight line, wherein this area may be the gap between the dashed lane lines, or the lane lines may be blurred due to weather reasons.

Step S24, calculate the confidence of each horizontal sliding window.

In order to determine whether a horizontal sliding window contains a lane line, the confidence of each horizontal sliding window needs to be calculated. Before calculating the confidence of each horizontal sliding window, the deep neural network model needs to be determined in advance. The deep neural network model is obtained through a large amount of sample data training. Lane line image samples under various road conditions and light conditions are obtained in advance, and the lane line primitives are labeled to obtain learning samples, so that the deep neural network can learn and determine the deep neural network model.

In an embodiment of the present application, a bird's-eye view including a horizontal sliding window is input into a preset deep learning neural network model, and the similarity between the primitive feature and the sample feature of each primitive in each horizontal sliding window is calculated. The confidence of each horizontal sliding window is determined according to the similarity. The similarity is proportional to the confidence. The higher the similarity, the higher the confidence. For example, if the similarity between a primitive feature in the foreground image and the corresponding sample feature is higher than the preset first ratio, for example, the first ratio is 98%, then the confidence of the primitive is recorded as 1. If the similarity between a primitive feature in the foreground image and the sample feature is lower than the preset second ratio, for example, the second ratio is 0.2%, then the confidence of the primitive is recorded as 0. According to the confidence of all primitives in the horizontal sliding window, the confidence of the horizontal sliding window is obtained.

FIG3 is a schematic diagram of the confidence provided by an embodiment of the present application. As shown in FIG3, assuming that there are 240 pixel points in the horizontal sliding window A whose pixel features meet the sample features, the confidence is 0.8, and assuming that there are 40 pixel points in the horizontal sliding window B whose pixel features meet the sample features, the confidence is 0.2.

Step S25, if there are a continuous preset number of horizontal sliding windows whose confidence is less than the preset threshold value, determine the previous horizontal sliding window of the continuous preset number of horizontal sliding windows as the previous horizontal sliding window, and determine the lane line end point of the main lane line according to the previous horizontal sliding window.

In an embodiment of the present application, after the confidence of each horizontal sliding window is calculated, the confidence is compared with the preset valve value. If the confidence of the continuous preset number of horizontal sliding windows is less than the preset valve value, the first horizontal sliding window in the continuous preset number of horizontal sliding windows is used as the previous horizontal sliding window.

Figure 4 is a schematic diagram of the confidence provided by an embodiment of this application.

For example, in some examples, assuming that the default valve value is 0.3 and the default quantity is 3, as shown in Figure 4, assuming that there are horizontal sliding windows 0, 2, 5, 7, 8, 9, and 10 whose confidences are all less than the default valve value 0.3, among which the confidences of the continuous horizontal sliding windows 7, 8, 9, and 10 are less than the default valve value 0.3, the previous horizontal sliding window (i.e., horizontal sliding window 6) of the continuous preset quantity (3) of horizontal sliding windows is taken as the previous horizontal sliding window.

After determining the first horizontal sliding window, based on the above confidence level, it can be determined that the sliding window moving after the first horizontal sliding window does not contain a lane line, and the lane line end point of the main lane line is determined according to the first horizontal sliding window (such as horizontal sliding window 6 in Figure 4).

In the embodiment of the present application, by calculating the confidence of the horizontal sliding window to determine whether the main lane line is included, not only the speed of simulating the lane line is accelerated, but also the accuracy of lane line recognition is improved.

Step S26, establish a vertical bar graph corresponding to the bird's-eye view, use the peak of the vertical bar graph as the starting point of the moving vertical sliding window, generate a target curve by fitting the non-zero primitive points in the vertical sliding window, and determine the target curve as a stop line if the end point of the lane line is on the target curve.

After obtaining the end point of the lane line, in order to further determine whether the position of the end point of the lane line is correct, a vertical bar graph corresponding to the bird's-eye view is established, and the target curve is determined using the vertical sliding window. By judging whether the end point of the lane line is on the target curve, it is determined whether the end point of the lane line is correct. Among them, the crest of the vertical sliding window is used as the starting point of moving the vertical sliding window, and the average value of the horizontal coordinates of all non-zero pixels in the previous vertical sliding window of the current vertical sliding window is calculated as the second horizontal coordinate average value. The vertical window center of the current vertical sliding window is determined according to the second horizontal coordinate average value, and the position of the current vertical sliding window is determined according to the vertical window center. The current vertical sliding window is moved to the target curve. Fit the non-zero primitive points in the vertical sliding window and all the vertical sliding windows before the current vertical sliding window to generate the target curve. If the non-zero primitive points in the current vertical sliding window are less than the preset non-zero primitive threshold value, the current vertical sliding window will not be fitted. That is, if there is no lane line primitive feature in the vertical sliding window, this vertical sliding window will not be fitted.

After obtaining the target curve, if the target curve and the end point of the lane line are consistent, it proves that there is a lane line at the end point of the lane line, that is, the stop line. If the target curve and the end point of the lane line are inconsistent, it proves that there is no lane line at the end point of the lane line. It is judged here that the end point of the lane line may be blurred due to external reasons such as weather.

Determining whether the positions of the target curve and the lane line end point are consistent may include: calculating the matching degree between the plurality of longitudinal sliding windows corresponding to the target curve and the previous transverse sliding windows, and if any matching degree exceeds a preset matching degree, determining that the lane line end point is located on the target curve. Determining the position of the stop line according to the position of the target curve and the position of the lane line end point. If the matching degree between the longitudinal sliding window moving on the stop line and the transverse sliding window is high, it can be further determined that the lane line end point obtained above is correct, and is not an inaccurate judgment due to blurry lane line representation, weak light or weather. Assuming that the target curve is not the stop line, the sliding window moving on the target curve has a low match with the horizontal sliding window, and it is judged that this is not the end of the lane line. The lane line may be blurred due to weather reasons, and a warning is output to remind the driver that the main lane line will not be recognized.

Figure 5 is a flow chart of the lane line detection method provided in an embodiment of this application.

In some examples, as shown in FIG5 , assuming that a vertical bar graph is established on the left side of the bird's-eye view, point C is the peak of the vertical bar graph, and point C is used as the starting point for moving the vertical sliding window on the right side of the bird's-eye view, and a target curve is simulated based on the moving vertical sliding window. If point C is located at the position of the previous horizontal sliding window (e.g., horizontal sliding window 6), it indicates that point C and the previous horizontal sliding window 6 are located on the stop line.

Figure 6 is a schematic diagram of lane lines provided in an embodiment of this application.

After the stop line is determined, a preset number of continuous horizontal sliding windows are filtered, such as horizontal sliding windows 7, 8, 9, and 10 in Figure 5. Horizontal sliding windows 7, 8, 9, and 10 are treated as noise and discarded, and do not participate in the lane line fitting process. As shown in Figure 6, the primitive points in the filtered horizontal sliding window are fitted to obtain the lane line including the end point of the lane line and the stop line. Compared with the actual lane line, the lane line simulated by discarding the noise has a higher matching degree with the actual lane line, which is convenient for assisted driving in the smart driving mode, improving driving safety and the driving experience of the driver.

In the embodiment of the present application, first, a foreground image of the vehicle is obtained by using a camera, and the foreground image is processed to convert the foreground image into a bird's-eye view, and a horizontal bar graph corresponding to the bird's-eye view is created. The peak of the horizontal bar graph is used as the starting point of the moving horizontal sliding window, and the main lane lines are generated by fitting the non-zero pixel points in the horizontal sliding window, including the left main lane line and the right main lane line. Then, the confidence of each horizontal sliding window is calculated. If there are a continuous preset number of horizontal sliding windows whose confidence is less than a preset threshold value, the previous horizontal sliding window of the continuous preset number of horizontal sliding windows is determined as the previous horizontal sliding window, and the lane line end point of the main lane line is determined according to the previous horizontal sliding window. Finally, in order to avoid the blurring of lane lines due to external factors such as weather, which leads to inaccurate identification of the end point of the lane line, a vertical bar graph corresponding to the bird's-eye view is established, and the peak of the vertical bar graph is used as the starting point of the moving vertical sliding window. The target curve is generated by fitting the non-zero primitive points in the vertical sliding window. If the end point of the lane line is located on the target curve, the target curve is the stop line. This application can improve the accuracy of identifying lane lines.

Please continue to refer to Figure 1. In this embodiment, the memory 11 can be an internal memory of the vehicle-mounted device 1, that is, a memory built into the vehicle-mounted device 1. In other embodiments, the memory 11 can also be an external memory of the vehicle-mounted device 1, that is, a memory externally connected to the vehicle-mounted device 1.

In some embodiments, the memory 11 is used to store program codes and various data, and to achieve high-speed and automatic access to programs or data during the operation of the vehicle-mounted device 1.

The memory 11 may include a random access memory, and may also include a non-volatile memory, such as a hard disk, a memory (Memory), a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a memory card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

In one embodiment, the processor 12 may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor, or the processor may be any other conventional processor, etc.

If the program code and various data in the memory 11 are implemented in the form of a software functional unit and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, such as the driving route planning method, which can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), etc.

It can be understood that the module division described above is a logical function division, and there may be other division methods in actual implementation. In addition, each functional module in each embodiment of the present application can be integrated in the same processing unit, or each module can exist physically separately, or two or more modules can be integrated in the same unit. The above-mentioned integrated module can be implemented in the form of hardware or in the form of hardware plus software functional modules.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of this application and are not limiting. Although this application is described in detail with reference to the preferred embodiments, ordinary technicians in this field should understand that the technical solution of this application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of this application.

S21~S26: Steps

Claims (10)

A lane line detection method, wherein the method includes: obtaining a foreground image of a vehicle; converting the foreground image into a bird's-eye view, and establishing a horizontal bar graph corresponding to the bird's-eye view; using the peak of the horizontal bar graph as the starting point of a moving horizontal sliding window, and generating a main lane line by fitting non-zero pixel points in the horizontal sliding window; calculating the confidence of each horizontal sliding window; if there are a preset number of continuous horizontal sliding windows whose confidences are all less than a preset threshold value, determining the lane line. The previous horizontal sliding window of a preset number of continuous horizontal sliding windows is used as the previous horizontal sliding window, and the lane end point of the main lane line is determined according to the previous horizontal sliding window; a vertical bar graph corresponding to the bird's-eye view is established, and the peak of the vertical bar graph is used as the starting point of the moving vertical sliding window. The target curve is generated by fitting the non-zero primitive points in the vertical sliding window. If the lane end point is located on the target curve, the target curve is determined to be a stop line. The lane line detection method as described in claim 1, wherein the converting the foreground image into a bird's-eye view comprises: performing distortion correction on the foreground image to obtain a corrected image; taking each non-zero pixel point in the corrected image as a target point, calculating the coordinates of the target point using a coordinate conversion formula to obtain an inverse perspective transformation matrix; and converting the corrected image into the bird's-eye view using the inverse perspective transformation matrix. The lane line detection method as claimed in claim 1, wherein the peak of the horizontal bar graph is used as the starting point for moving the horizontal sliding window, and the main lane line is generated by fitting the non-zero pixel points in the horizontal sliding window, including: using the peak of the horizontal bar graph as the starting point for moving the horizontal sliding window, calculating the horizontal position of all non-zero pixels in the previous horizontal sliding window of the current horizontal sliding window, and generating a main lane line. The average value of the coordinates is used as the first horizontal coordinate average value; the horizontal window center of the current horizontal sliding window is determined according to the first horizontal coordinate average value, and the position of the current horizontal sliding window is determined according to the horizontal window center; the non-zero primitive points in the current horizontal sliding window and all horizontal sliding windows before the current horizontal sliding window are fitted to generate the main lane line. The lane line detection method as described in claim 1, wherein the calculating the confidence of each of the horizontal sliding windows comprises: inputting the bird's-eye view containing the horizontal sliding window into a preset deep learning neural network model, calculating the similarity between the primitive features and the sample features of each primitive in each of the horizontal sliding windows; determining the confidence of each of the horizontal sliding windows according to the similarity, wherein the similarity is proportional to the confidence. The lane line detection method as claimed in claim 1, wherein the peak of the vertical bar graph is used as the starting point for moving the vertical sliding window, and the target curve is generated by fitting the non-zero pixel points in the vertical sliding window, including: using the peak of the vertical sliding window as the starting point for moving the vertical sliding window, calculating the number of all non-zero pixels in the previous vertical sliding window of the current vertical sliding window, and calculating the number of all non-zero pixels in the previous vertical sliding window of the current vertical sliding window. The average value of the horizontal coordinate is used as the second horizontal coordinate average value; the vertical window center of the current vertical sliding window is determined according to the second horizontal coordinate average value, and the position of the current vertical sliding window is determined according to the vertical window center; the non-zero primitive points in the current vertical sliding window and all vertical sliding windows before the current vertical sliding window are fitted to generate the target curve. The lane line detection method as described in claim 1, wherein if the end point of the lane line is located on the target curve, determining that the target curve is a stop line includes: calculating the matching degree between a plurality of longitudinal sliding windows corresponding to the target curve and the previous transverse sliding window; if any matching degree exceeds a preset matching degree, determining that the end point of the lane line is located on the target curve; and determining the position of the stop line based on the position of the target curve and the position of the end point of the lane line. The lane line detection method as described in claim 6, wherein after determining that the target curve is a stop line, the method further comprises: filtering the preset number of continuous horizontal sliding windows. The lane line detection method as described in claim 1, wherein the method further comprises: if the lane line end point does not exist on the target curve, determining that the stop line does not exist in the foreground image, and outputting a warning to indicate that the main lane line will not be recognized. A vehicle-mounted device, wherein the vehicle-mounted device includes a processor and a memory, and the processor is used to execute a computer program stored in the memory to implement the lane line detection method as described in any one of claims 1 to 8. A computer-readable storage medium, wherein the computer-readable storage medium stores at least one instruction, and when the at least one instruction is executed by a processor, the lane line detection method as described in any one of claims 1 to 8 is implemented.

Images