A hole leakage-proof visual detection method and system for automobile manufacturing

Abstract

The application discloses a kind of leak-proof hole visual inspection method and system for automobile manufacturing, it is related to industrial automation technical field, including: based on binary image, respectively in each ellipse detection area Statistics white pixel number, calculate the proportion of white pixel in total pixel of corresponding ellipse detection area, obtain white pixel proportion, compare the white pixel proportion corresponding to each ellipse detection area with preset hole site determination threshold, obtain hole site determination result;Based on the hole site determination result corresponding to each ellipse detection area, the detection results of all industrial cameras and all ellipse detection areas are summarized, and the overall detection result of workpiece is output, and the overall detection result of workpiece is used to control production line to execute sorting and alarm operation.The application improves the accuracy, automation degree and quality control stability of leak-proof hole detection through Gaussian filter denoising and brightness equalization processing.

Description

Technical Field

[0001] This invention relates to the field of industrial automation technology, and in particular to a visual inspection method and system for preventing leaks in automobile manufacturing. Background Technology

[0002] In automotive parts manufacturing, laser cutting technology is widely used for hole processing and contour cutting of metal sheets. To improve inspection efficiency, the industry is gradually adopting industrial cameras and image processing technology to automatically detect laser-cut holes. By performing grayscale segmentation, edge extraction, or pixel statistics on the hole position image, the presence or absence of holes can be determined. In some production lines, PLC control can be used to provide feedback on the inspection results, allowing for automatic sorting or stopping the production line for manual intervention.

[0003] In actual production, workpiece surfaces often exhibit reflectivity, slag, oil stains, and uneven local lighting, making visual inspection methods based on fixed thresholds or uniform parameters prone to false positives or false negatives. Furthermore, traditional inspection solutions typically rely on a single camera or fixed parameters, making it difficult to simultaneously meet the demands of large-size workpieces, multi-hole locations, high precision, and high-speed online inspection. This also hinders the flexible production requirements of diverse parts, thus impacting inspection reliability and production efficiency. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a visual inspection method for leak-proof holes in automobile manufacturing to solve the problems of low inspection efficiency and untimely online sorting feedback caused by workpiece reflection and the dispersion of multiple holes.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a visual inspection method for leak-proof holes in automobile manufacturing, comprising: activating multiple industrial cameras to simultaneously expose and acquire images of the workpiece to be inspected, thereby obtaining images of the holes to be inspected; sequentially performing Gaussian filtering noise reduction and brightness equalization processing on the images of the holes to be inspected to obtain preprocessed images; defining multiple elliptical detection regions based on the preprocessed images, performing regionalized adaptive threshold segmentation on each region, and generating binarized images by setting block size parameters and threshold compensation parameters for each elliptical detection region; based on the binarized images, counting the number of white pixels in each elliptical detection region, calculating the proportion of white pixels to the total pixels in the corresponding elliptical detection region, obtaining the white pixel percentage, comparing the white pixel percentage corresponding to each elliptical detection region with a preset hole position determination threshold, and obtaining the hole position determination result; summarizing the detection results of all industrial cameras and all elliptical detection regions based on the hole position determination results of each elliptical detection region, outputting the overall workpiece inspection result, and controlling the production line to perform sorting and alarm operations based on the overall workpiece inspection result.

[0007] As a preferred embodiment of the visual inspection method for leak-proof holes in automobile manufacturing described in this invention, the image of the hole to be inspected is obtained by a drive motor driving a turntable to rotate via a reducer, a clamping device fixing the workpiece to be inspected, and transporting the workpiece to be inspected to the inspection station, and activating multiple industrial cameras to simultaneously expose and acquire images of the workpiece to be inspected.

[0008] As a preferred embodiment of the visual inspection method for leak-proof holes in automobile manufacturing described in this invention, the image of the hole to be inspected is sequentially subjected to Gaussian filtering for noise reduction and brightness equalization to obtain a preprocessed image. The specific steps are as follows: Gaussian filtering is applied to the image of the hole to be detected to remove random noise, resulting in the Gaussian-filtered image of the hole to be detected. The image of the hole to be detected after Gaussian filtering is subjected to brightness equalization processing to adjust the local gray-level distribution in the image of the hole to be detected, thus obtaining the preprocessed image.

[0009] As a preferred embodiment of the visual inspection method for leak-proof holes in automobile manufacturing described in this invention, the elliptical detection area is established by drawing, dragging and scaling with a mouse, and corresponds to the target hole area in the workpiece to be inspected. The block size parameter and threshold compensation parameter are set according to the target hole size, ellipse detection area size, grayscale difference between hole foreground and surrounding background, local reflectivity and illumination uniformity of each ellipse detection area.

[0010] As a preferred embodiment of the visual inspection method for leak-proof holes in automobile manufacturing described in this invention, the specific steps for generating the binarized image are as follows: Based on the block size parameter corresponding to each elliptical detection region, the local grayscale mean value within the neighborhood of the current pixel is obtained; The local threshold of the current pixel is calculated based on the local gray-level mean and threshold compensation parameters. Then, the local threshold of the current pixel is used to binarize each pixel in the elliptical detection area to generate a binarized image.

[0011] In a preferred embodiment of the visual inspection method for preventing leaks in automobile manufacturing described in this invention, the expression for calculating the local threshold of the current pixel is: ; in, Represented in pixels The local grayscale mean value centered at a block size equal to the size of its neighborhood. Represents an adjustable constant. Represents pixels Local threshold, This represents the column coordinate of a pixel in the image coordinate system. This represents the row coordinate of a pixel in the image coordinate system.

[0012] As a preferred embodiment of the visual inspection method for leak-proof holes in automobile manufacturing described in this invention, the hole position determination result includes hole position qualified and hole position unqualified; when the proportion of white pixels exceeds the hole position determination threshold, the corresponding hole position is determined to be qualified, and when the proportion of white pixels does not exceed the hole position determination threshold, the corresponding hole position is determined to be unqualified.

[0013] As a preferred embodiment of the visual inspection method for leak-proof holes in automobile manufacturing described in this invention, the hole position determination threshold is predetermined based on the target hole position size, the area of ​​the elliptical detection region, the proportional relationship between the light-transmitting imaging area of ​​the hole and the area of ​​the elliptical detection region, and the allowable imaging deviation of the target hole position.

[0014] As a preferred embodiment of the visual inspection method for preventing leaks in automobile manufacturing described in this invention, the specific steps of controlling the production line to perform sorting and alarm operations based on the overall inspection results of the workpiece are as follows: The results of hole position determination for all industrial cameras and all elliptical detection areas are summarized to generate the overall workpiece inspection results; When the overall inspection result of the workpiece is qualified, the PLC controls the workpiece to be inspected to enter the next process; when the overall inspection result of the workpiece is unqualified, the PLC controls the production line to perform the corresponding sorting operation and outputs alarm information.

[0015] Secondly, this invention provides a visual inspection system for leak-proof holes in automobile manufacturing, comprising: an image acquisition module for activating multiple industrial cameras to simultaneously expose and acquire images of the workpiece to be inspected, thereby obtaining images of the holes to be inspected; an image preprocessing module for sequentially performing Gaussian filtering noise reduction and brightness equalization on the images of the holes to be inspected, thereby obtaining a preprocessed image; a detection area configuration module for defining multiple elliptical detection areas based on the preprocessed image, performing regionalized adaptive threshold segmentation on each, and setting block size parameters and threshold compensation parameters according to each elliptical detection area to generate a binarized image; a hole detection module for counting the number of white pixels in each elliptical detection area based on the binarized image, calculating the proportion of white pixels to the total pixels of the corresponding elliptical detection area, obtaining the white pixel percentage, comparing the white pixel percentage corresponding to each elliptical detection area with a preset hole position determination threshold, thereby obtaining a hole position determination result; and a result control module for summarizing the detection results of all industrial cameras and all elliptical detection areas based on the hole position determination results of each elliptical detection area, outputting the overall workpiece inspection result, and controlling the production line to perform sorting and alarm operations based on the overall workpiece inspection result.

[0016] The beneficial effects of this invention are as follows: by simultaneously exposing and acquiring images of the holes to be inspected using multiple industrial cameras, the online inspection efficiency of large-sized, multi-hole automotive parts is improved; and by using Gaussian filtering for noise reduction and brightness equalization processing, the influence of random noise, local reflections, shadows, and uneven lighting on the grayscale distribution of the holes is reduced, and the production line is controlled to perform sorting and alarm operations, thereby improving the accuracy, automation level, and quality control stability of leak-proof hole detection. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a visual inspection method for preventing leaks in automobile manufacturing.

[0019] Figure 2 This is a schematic diagram of a vision inspection system for preventing leaks in automobile manufacturing.

[0020] Figure 3 The flowchart shows the preprocessing steps for the image of the borehole to be detected.

[0021] Figure 4 This is a flowchart for hole location determination and production line control. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a visual inspection method for leak-proof holes in automobile manufacturing, comprising the following steps: S1. Start multiple industrial cameras to simultaneously expose and acquire images of the workpiece to be inspected, obtaining images of the holes to be inspected.

[0026] The image of the hole to be inspected is obtained by driving a turntable to rotate via a drive motor and a reducer, fixing the workpiece to be inspected with a clamping device, transporting the workpiece to the inspection station, and activating multiple industrial cameras to simultaneously expose and collect images of the workpiece.

[0027] Specifically, the drive motor outputs rotational power, and the reducer reduces the speed of the drive motor output before transmitting it to the turntable. The turntable rotates under the drive of the reducer, and the clamping device fixes the position of the workpiece to be inspected to prevent the workpiece from shifting during the rotation of the turntable. After the workpiece to be inspected is transported to the inspection station by the turntable, multiple industrial cameras synchronously perform exposure and acquisition according to a preset trigger sequence. The multiple industrial cameras acquire images corresponding to different areas of the hole to be inspected, resulting in images of the hole to be inspected covering multiple areas of the hole to be inspected.

[0028] S2. The image of the hole to be detected is sequentially subjected to Gaussian filtering for noise reduction and brightness equalization to obtain a preprocessed image.

[0029] Gaussian filtering is applied to the image of the hole to be inspected to remove random noise, resulting in the Gaussian-filtered image of the hole to be inspected.

[0030] Specifically, based on the images of the holes to be inspected acquired by simultaneous exposure of multiple industrial cameras, Gaussian filtering is applied to the images. The Gaussian filtering process uses a Gaussian convolution kernel to perform neighborhood weighted averaging on the pixel gray values ​​in the images of the holes to be inspected, and smooths out discrete bright spots, random noise, and local gray-level abrupt change areas in the images of the holes to be inspected, thereby reducing random noise interference in the images of the holes to be inspected, while preserving the overall contour information and gray-level distribution characteristics of the area of ​​the holes to be inspected, thus obtaining the images of the holes to be inspected after Gaussian filtering.

[0031] The image of the hole to be detected after Gaussian filtering is subjected to brightness equalization processing to adjust the local gray-level distribution in the image of the hole to be detected, thus obtaining the preprocessed image.

[0032] Specifically, based on the Gaussian-filtered image of the hole to be inspected, brightness equalization processing is performed on the Gaussian-filtered image of the hole to be inspected. Brightness equalization processing readjusts the gray-level distribution in the image of the hole to be inspected, improves the gray-level contrast between the hole to be inspected area and the background area, and performs gray-level balancing processing on the locally overbright and locally underbright areas in the image of the hole to be inspected. This reduces the influence of reflections, shadows and uneven local lighting on the surface of the workpiece to be inspected on the gray-level distribution of the image of the hole to be inspected, and enhances the gray-level change characteristics at the edges of the hole to be inspected area, thus obtaining the preprocessed image.

[0033] S3. Based on the preprocessed image, define multiple elliptical detection regions, perform regionalized adaptive threshold segmentation on each region, and set block size parameters and threshold compensation parameters according to each elliptical detection region to generate a binarized image.

[0034] The elliptical detection area is created by drawing, dragging, and scaling with the mouse, and corresponds to the target hole area in the workpiece to be inspected.

[0035] Specifically, the preprocessed image displays the hole position imaging of the workpiece to be inspected. Based on the position of the target hole position area, an elliptical detection area is formed by drawing with the mouse. The position of the elliptical detection area is adjusted by dragging with the mouse, and the major and minor axis dimensions of the elliptical detection area are adjusted by scaling with the mouse, so that the elliptical detection area covers the corresponding target hole position area and establishes a correlation with the corresponding target hole position area for subsequent independent image processing of the target hole position area.

[0036] The block size parameter and threshold compensation parameter are set according to the target hole size, ellipse detection area size, grayscale difference between hole foreground and surrounding background, local reflectivity and illumination uniformity of each ellipse detection area.

[0037] Based on the block size parameter corresponding to each elliptical detection region, the local grayscale mean value within the neighborhood of the current pixel is obtained.

[0038] Specifically, taking the current pixel as the center, a square neighborhood is determined based on the block size parameter of the corresponding elliptical detection region. The block size parameter is the side length of the square neighborhood and must be an odd number. The gray values ​​of each pixel within the square neighborhood are obtained, and a weighted average is calculated on the gray values ​​of each pixel within the square neighborhood to obtain the local gray mean within the neighborhood of the current pixel. Among them, the closer the pixel is to the current pixel, the greater the weight, and the farther the pixel is from the current pixel, the smaller the weight, thereby reducing the impact of local noise and edge abrupt changes on the local gray mean.

[0039] The local threshold of the current pixel is calculated based on the local gray-level mean and threshold compensation parameters. Then, the local threshold of the current pixel is used to binarize each pixel in the elliptical detection area to generate a binarized image.

[0040] It should be noted that the expression for calculating the local threshold of the current pixel is: ; in, Represented in pixels The local grayscale mean value centered at a block size equal to the size of its neighborhood. Represents an adjustable constant. Represents pixels Local threshold, This represents the column coordinate of a pixel in the image coordinate system. This represents the row coordinate of a pixel in the image coordinate system.

[0041] S4. Based on the binarized image, count the number of white pixels in each elliptical detection region, calculate the proportion of white pixels to the total number of pixels in the corresponding elliptical detection region, obtain the proportion of white pixels, compare the proportion of white pixels in each elliptical detection region with the preset hole position determination threshold, and obtain the hole position determination result.

[0042] The hole position determination result includes hole position qualified and hole position unqualified; when the proportion of white pixels exceeds the hole position determination threshold, the corresponding hole position is determined to be qualified, and when the proportion of white pixels does not exceed the hole position determination threshold, the corresponding hole position is determined to be unqualified.

[0043] Specifically, pixel statistics are performed on the binarized image corresponding to each ellipse detection region to obtain the number of white pixels and the total number of pixels in the ellipse detection region. The proportion of white pixels is obtained based on the ratio of the number of white pixels to the total number of pixels. The proportion of white pixels is compared with the hole position determination threshold of the corresponding ellipse detection region. When the proportion of white pixels exceeds the hole position determination threshold, the corresponding hole position is determined to be qualified. When the proportion of white pixels does not exceed the hole position determination threshold, the corresponding hole position is determined to be unqualified.

[0044] The hole position determination threshold is predetermined based on the target hole position size, the area of ​​the elliptical detection region, the ratio between the light-transmitting imaging area of ​​the hole and the area of ​​the elliptical detection region, and the allowable imaging deviation of the target hole position.

[0045] Furthermore, based on the target hole size, the light-transmitting imaging area of ​​the hole when it is normally transparent in the preprocessed image is determined, and the area of ​​the elliptical detection region is determined based on the major and minor axis dimensions of the elliptical detection region; then, the proportion of the light-transmitting imaging area of ​​the hole in the area of ​​the elliptical detection region is calculated to obtain the theoretical proportion of white pixels when the target hole is normally present. Based on the influence of the workpiece positioning deviation, industrial camera imaging deviation, hole edge burrs, surface reflection of the workpiece, and local oil stains on the light transmission imaging area of ​​the hole, an imaging deviation range is reserved for the theoretical white pixel ratio. The lower limit of the white pixel ratio after the reserved imaging deviation is used as the preset hole position determination threshold of the corresponding elliptical detection area. The hole position determination threshold can match the target hole size and the area of ​​the elliptical detection area, and is used to determine whether the white pixel ratio in the elliptical detection area meets the conditions for the existence of the hole.

[0046] S5. Based on the hole position determination results of each elliptical detection area, summarize the detection results of all industrial cameras and all elliptical detection areas, output the overall workpiece detection result, and control the production line to perform sorting and alarm operations based on the overall workpiece detection result.

[0047] The results of hole position determination for all industrial cameras and all elliptical detection areas are summarized to generate the overall inspection results of the workpiece.

[0048] Specifically, based on the hole position determination results corresponding to each elliptical detection area, the hole position determination results of all industrial cameras and all elliptical detection areas are summarized. By statistically analyzing the pass and fail status of all holes on each workpiece to be inspected, it is determined whether there are any holes on the workpiece to be inspected that do not meet the preset hole position determination threshold. If all holes meet their respective preset hole position determination thresholds, the workpiece to be inspected is determined to be qualified as a whole; otherwise, the workpiece to be inspected is determined to be unqualified as a whole, thereby generating the overall inspection result of the workpiece.

[0049] When the overall inspection result of the workpiece is qualified, the PLC controls the workpiece to be inspected to enter the next process; when the overall inspection result of the workpiece is unqualified, the PLC controls the production line to perform the corresponding sorting operation and outputs alarm information.

[0050] This embodiment also provides a visual inspection system for leak-proof holes in automobile manufacturing, including: an image acquisition module for activating multiple industrial cameras to simultaneously expose and acquire images of the workpiece to be inspected, thereby obtaining images of the holes to be inspected; an image preprocessing module for sequentially performing Gaussian filtering noise reduction and brightness equalization on the images of the holes to be inspected, thereby obtaining a preprocessed image; a detection area configuration module for defining multiple elliptical detection areas based on the preprocessed image, performing regionalized adaptive threshold segmentation on each, and setting block size parameters and threshold compensation parameters according to each elliptical detection area to generate a binarized image; a hole detection module for counting the number of white pixels in each elliptical detection area based on the binarized image, calculating the proportion of white pixels to the total pixels in the corresponding elliptical detection area, obtaining the white pixel percentage, comparing the white pixel percentage corresponding to each elliptical detection area with a preset hole position determination threshold, and obtaining a hole position determination result; and a result control module for summarizing the detection results of all industrial cameras and all elliptical detection areas based on the hole position determination results of each elliptical detection area, outputting the overall workpiece inspection result, and controlling the production line to perform sorting and alarm operations based on the overall workpiece inspection result.

[0051] In summary, this invention improves the online inspection efficiency of large-sized, multi-hole automotive parts by simultaneously exposing and acquiring images of the holes to be inspected using multiple industrial cameras; and by reducing noise through Gaussian filtering and brightness equalization processing, it reduces the impact of random noise, local reflections, shadows, and uneven lighting on the grayscale distribution of the holes, controls the production line to perform sorting and alarm operations, and improves the accuracy, automation, and quality control stability of leak-proof hole detection.

[0052] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A visual inspection method for leak-proof holes in automobile manufacturing, characterized in that, include: Multiple industrial cameras are activated to simultaneously expose and capture images of the workpiece to be inspected, thereby obtaining images of the holes to be inspected. The images of the holes to be inspected are sequentially subjected to Gaussian filtering for noise reduction and brightness equalization to obtain preprocessed images; Based on the preprocessed image, multiple elliptical detection regions are defined, and regionalized adaptive threshold segmentation is performed on each region. Block size parameters and threshold compensation parameters are set according to each elliptical detection region to generate a binarized image. The elliptical detection area is created by drawing, dragging, and scaling with a mouse, and corresponds to the target hole area in the workpiece to be detected. The block size parameter and threshold compensation parameter are set according to the target hole size, ellipse detection area size, grayscale difference between hole foreground and surrounding background, local reflectivity and illumination uniformity of each ellipse detection area. The specific steps for generating the binarized image are as follows: Based on the block size parameter corresponding to each elliptical detection region, the local grayscale mean value within the neighborhood of the current pixel is obtained; The local threshold of the current pixel is calculated based on the local gray-level mean and threshold compensation parameters, and the local threshold of the current pixel is used to perform binarization processing on each pixel in the elliptical detection area to generate a binarized image. Based on the binarized image, the number of white pixels is counted in each elliptical detection region, and the proportion of white pixels to the total pixels in the corresponding elliptical detection region is calculated to obtain the white pixel proportion. The white pixel proportion corresponding to each elliptical detection region is compared with the preset hole position determination threshold to obtain the hole position determination result. The expression for calculating the local threshold of the current pixel is: ； in, Represented in pixels The local grayscale mean value centered at a block size equal to the size of its neighborhood. Represents an adjustable constant. Represents pixels Local threshold, This represents the column coordinate of a pixel in the image coordinate system. This represents the row coordinate of a pixel in the image coordinate system; Based on the hole position determination results of each elliptical detection area, the detection results of all industrial cameras and all elliptical detection areas are summarized, and the overall workpiece detection result is output. Based on the overall workpiece detection result, the production line is controlled to perform sorting and alarm operations.

2. The visual inspection method for leak-proof holes in automobile manufacturing as described in claim 1, characterized in that, The image of the hole to be inspected is obtained by driving a turntable to rotate via a drive motor and a reducer, fixing the workpiece to be inspected with a clamping device, transporting the workpiece to the inspection station, and activating multiple industrial cameras to simultaneously expose and collect images of the workpiece.

3. The visual inspection method for leak-proof holes in automobile manufacturing as described in claim 2, characterized in that, The image of the hole to be detected is sequentially subjected to Gaussian filtering for noise reduction and brightness equalization to obtain a preprocessed image. The specific steps are as follows: Gaussian filtering is applied to the image of the hole to be detected to remove random noise, resulting in the Gaussian-filtered image of the hole to be detected. The image of the hole to be detected after Gaussian filtering is subjected to brightness equalization processing to adjust the local gray-level distribution in the image of the hole to be detected, thus obtaining the preprocessed image.

4. The visual inspection method for leak-proof holes in automobile manufacturing as described in claim 1, characterized in that, The hole position determination result includes hole position qualified and hole position unqualified; when the proportion of white pixels exceeds the hole position determination threshold, the corresponding hole position is determined to be qualified, and when the proportion of white pixels does not exceed the hole position determination threshold, the corresponding hole position is determined to be unqualified.

5. The visual inspection method for leak-proof holes in automobile manufacturing as described in claim 1, characterized in that, The aperture position determination threshold is determined based on the target aperture size, the area of ​​the elliptical detection region, the ratio between the light-transmitting imaging area of ​​the aperture and the area of ​​the elliptical detection region, and the allowable imaging deviation of the target aperture.

6. The visual inspection method for leak-proof holes in automobile manufacturing as described in claim 5, characterized in that, The specific steps for controlling the production line to perform sorting and alarm operations based on the overall workpiece inspection results are as follows: The results of hole position determination for all industrial cameras and all elliptical detection areas are summarized to generate the overall workpiece inspection results; When the overall inspection result of the workpiece is qualified, the PLC controls the workpiece to be inspected to enter the next process; when the overall inspection result of the workpiece is unqualified, the PLC controls the production line to perform the corresponding sorting operation and outputs alarm information.

7. A visual inspection system for leak-proof holes in automobile manufacturing, based on the visual inspection method for leak-proof holes in automobile manufacturing according to any one of claims 1 to 6, characterized in that, include: The image acquisition module is used to activate multiple industrial cameras to simultaneously expose and acquire images of the workpiece to be inspected, thereby obtaining images of the holes to be inspected. The image preprocessing module is used to sequentially perform Gaussian filtering noise reduction and brightness equalization processing on the image of the hole to be detected to obtain a preprocessed image; The detection region configuration module is used to define multiple elliptical detection regions based on the preprocessed image, perform regionalized adaptive threshold segmentation on each region, and set block size parameters and threshold compensation parameters according to each elliptical detection region to generate a binarized image. The hole location detection module is used to count the number of white pixels in each elliptical detection region based on the binarized image, calculate the proportion of white pixels to the total pixels in the corresponding elliptical detection region, obtain the white pixel proportion, and compare the white pixel proportion of each elliptical detection region with the preset hole location determination threshold to obtain the hole location determination result. The result control module is used to summarize the detection results of all industrial cameras and all elliptical detection areas based on the hole position determination results of each elliptical detection area, output the overall workpiece detection result, and control the production line to perform sorting and alarm operations based on the overall workpiece detection result.

Images