Quality Inspection System Based on Computer Vision

Abstract

The present invention discloses a quality inspection system based on computer vision, including an image acquisition module, a quality inspection module and a defect processing module, characterized in that: the image acquisition module is used to collect the surface visual image of the processed metal workpiece, collect the metal workpiece detection data and enter the various parameter standards of the metal workpiece into the system, the quality inspection module is used to detect the edge burrs, surface scratches and internal cracks of the metal workpiece, the defect processing module is used to process the detected edge burrs and polish the processed edges, the image acquisition module. The quality inspection module and the defect processing module are electrically connected to each other, the image acquisition module includes a visual module, a sensor module and a product information entry module, the visual module is used to shoot the visual image of the metal workpiece during the processing, and the present invention has the characteristics of high detection accuracy and high efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of computer vision, and in particular to a quality inspection system based on computer vision. Background Art

[0002] In recent years, with the development of science and technology in my country, the manufacturing industry has gradually become intelligent. The existing technology uses molds and presses to stamp and forge metal materials to make them plastically deform or separate to obtain processed workpieces. The existing quality inspection methods for metal stamping parts mostly use a sample comparison method that compares the model of the stamping part with the sample model. The processing environment of metal stamping parts is complex, and there are also differences in the processed workpieces. Existing metal workpieces will have defects such as burrs, roller marks, cracks and scratches during the processing. However, when the existing technology detects defects in metal workpieces, during the processing of metal workpieces, the different metal materials cause different reflection intensities, resulting in different textures on the surface of the workpiece. The changes lead to certain errors in the defect detection at the reflective part. Moreover, when the camera is shooting, due to the different sizes of the workpieces, the imaging of the visual images of the workpieces taken by the industrial camera is also different, resulting in the inability to match the edge detection features, misjudgment and missed judgment, resulting in low accuracy of the defect detection results. Moreover, the existing technology uses the method of compensating for the detected unqualified products by relying on the experience of the experienced craftsman to judge whether the workpiece can be remedied and formulate a remedial plan. Due to the wide variety of metal workpieces, the process of workpiece remediation is slow, which in turn affects the quality inspection efficiency of metal workpieces. Therefore, it is very necessary to design a quality inspection system based on computer vision with high detection accuracy and efficiency. Summary of the invention

[0003] The purpose of the present invention is to provide a quality inspection system based on computer vision to solve the problems raised in the above background technology.

[0004] In order to solve the above technical problems, the present invention provides the following technical solutions: a quality inspection system based on computer vision, comprising an image acquisition module, a quality inspection module and a defect processing module, wherein the image acquisition module is used to collect surface visual images of processed metal workpieces, collect metal workpiece detection data and enter various parameter standards of the metal workpiece into the system, the quality inspection module is used to detect edge burrs, surface scratches and internal cracks of the metal workpiece, the defect processing module is used to process the detected edge burrs and polish the processed edges, and the image acquisition module, the quality inspection module and the defect processing module are electrically connected to each other;

[0005] The image processing module includes a reflection processing submodule and an image normalization processing submodule, wherein the reflection processing submodule is used to process the reflection part of the photographed metal workpiece, and the image normalization processing submodule is used to perform image normalization processing on the collected visual image of the metal workpiece;

[0006] The defect detection module includes a burr detection submodule, a scratch detection submodule and a crack detection submodule. The burr detection submodule is used to detect whether there are burrs on the edge of the metal workpiece, the scratch detection submodule is used to detect whether there are scratches on the metal workpiece, and the crack detection submodule is used to detect whether there are cracks inside the metal workpiece.

[0007] According to the above technical solution, the image acquisition module includes a visual module, a sensor module and a product information entry module. The visual module is used to capture the visual image of the metal workpiece during the processing process, the sensor module is used to measure the edge size of the processed metal workpiece, and the product information entry module is used to enter the comprehensive parameter data of the product produced by the current stamping machine into the system.

[0008] According to the above technical solution, the quality detection module includes an image processing module, which is used to identify shadows and reflections in the image, and process the shadow areas and reflection areas in the image to reduce the impact of shadows and reflections on quality detection.

[0009] According to the above technical solution, the defect processing module includes a burr processing module, an edge analysis module and a polishing module. The burr processing module is used to analyze the size, thickness and connection area of ​​the identified burrs, and process the edge burrs of the metal workpiece according to the analysis results. The edge analysis module is used to analyze the cross-section after cutting and analyze the residue of the incision after cutting the burrs. The polishing module is used to polish the workpiece cross-section according to the comprehensive parameters of the metal workpiece.

[0010] According to the above technical solution, the quality detection module also includes a defect detection module, and the defect detection module is used to perform quality inspection on the processed metal workpiece.

[0011] According to the above technical solution, the operation method of the quality inspection system mainly includes the following steps:

[0012] Step S1: Using the visual module, a visual image of the stamped metal workpiece on the conveyor belt is captured in real time and uploaded to the system; using the sensor module, the comprehensive detection data collected by the sensor is uploaded to the system; and using the product information entry module, the detailed parameter data of the metal workpiece is entered into the system;

[0013] Step S2: After the data collection is completed, the system starts the image processing module, starts to detect whether there is reflection in the image, and processes the reflection part in the image;

[0014] Step S3: When detecting whether the metal workpiece has defects, the system sends an electrical signal to start the defect detection module, starts analyzing the defects of the metal workpiece, and marks the metal workpiece according to the analysis results;

[0015] Step S4: When processing defects on a metal workpiece, start the burr processing module, start analyzing various parameters of the burr, analyze the burr processing method, and process the burr according to the burr analysis results.

[0016] According to the above technical solution, step S2 further includes the following steps:

[0017] Step S21: Retrieving a visual image of the metal workpiece, scanning and identifying the contour nodes of the punching table, cropping the visual image according to the identified contour nodes of the punching table, and normalizing the cropped visual image using a weighted median filtering method;

[0018] Step S22: Segment the processed image in equal proportions, identify the image features in each image block, including color features and brightness features, and calculate the standard deviation of the color values ​​of the pixels in the current image block using the formula Where i = 1, 2, 3...n, F represents the standard deviation of the pixel color value in the current image block, X i Represents the color value of each pixel. It represents the average color value of the metal workpiece surface set by the system. If the standard deviation F of the pixel color value in the current image block is greater than the threshold set by the system, the current image block is marked as a reflective area image block, otherwise the current image block is marked as a non-reflective area image block. The mark in the image block is read, the image blocks with the same mark are merged, and the reflective area is identified and marked.

[0019] According to the above technical solution, step S3 further includes the following steps:

[0020] Step S31: retrieve the visual image of the metal workpiece, grayscale the metal workpiece image, identify the grayscale values ​​of all pixels in the current metal workpiece image, record and save it in a collection The value of the enhanced pixel brightness corresponding to the gray value change in the X-axis direction is calculated by the formula S = 2G mn -G m(n+1) , the value of the enhanced pixel brightness corresponding to the change in grayscale value in the Y-axis direction is calculated by the formula W = 2G mn -G (m-1)n, if the value S or W of the enhanced pixel brightness corresponding to the gray value change is greater than the maximum threshold set by the system, the current pixel is marked as a scratch pixel, otherwise the current pixel is removed;

[0021] Step S32: Identify edge feature nodes of the metal workpiece image, mark the first edge of the current metal workpiece according to the edge feature nodes of the metal workpiece, identify the color value of the first edge pixel of the metal workpiece, and when the difference between the color value of the edge pixel of the metal workpiece and the color value of the adjacent pixel is greater than the threshold set by the system, mark the current pixel as an edge pixel, retrieve the color value of the edge pixel, and compare the color value of the edge pixel with the threshold set by the system. If the color value of the current edge pixel is within the threshold range set by the system, retain the mark of the current edge pixel, otherwise delete the edge pixel mark of the current pixel, retrieve the marked edge pixel, connect the adjacent edge pixels, and construct an edge model of the metal workpiece;

[0022] Step S33: retrieve the visual image of the metal workpiece, scan and anchor the feature points on the metal workpiece, establish a coordinate system in the image, and calculate the distance between the edge pixel point in the current metal workpiece image and the feature point of the metal workpiece by using a formula Wherein, L represents the distance between the current edge pixel point and the feature point of the metal workpiece, (X1, Y1) represents the coordinates of the current edge pixel point, (X2, Y2) represents the coordinates of the feature node of the metal workpiece, if the distance L between the edge pixel point in the current metal workpiece image and the feature point of the metal workpiece does not match the system set threshold, the current edge pixel point is marked as a burr, otherwise the system continues to detect, identify the edge pixel point marked with the burr, and calculate the distance L2 between the adjacent edge pixel points marked with the burr, if L2 is less than the system set threshold, the two adjacent edge pixels are connected to construct a burr model, otherwise the system continues to detect.

[0023] According to the above technical solution, in step S31, the marked scratch pixel points are retrieved, and it is identified whether the scratch pixel points are reflective pixels. If there is a reflective mark in the current scratch pixel point, the scratch mark in the current scratch pixel point is deleted. Otherwise, the scratch pixel points are connected to construct a scratch model, and the color value of the pixel with the largest pixel color value in the scratch model is measured. If the color value of this pixel point is less than the system set threshold, the current pixel point color value is marked. Otherwise, the system continues to detect and retrieve ultrasonic detection data, and compares the ultrasonic detection data with the database. If the similarity is less than the threshold, it is marked that the current metal material has cracks, otherwise there are no cracks.

[0024] According to the above technical solution, in step S4, the burr detection data of the metal workpiece is retrieved, the anchor is located at the burr edge node on the edge of the metal workpiece, the edge turning angle of the metal workpiece between the two current burr edge nodes is identified, and the length L3 that needs to be cut of the current burr is calculated according to the formula. If the current cutting length L3 is greater than the threshold set by the system, the current metal workpiece is marked as scrapped, otherwise the current metal workpiece is marked as a remediable workpiece, and the remediable workpiece is retrieved and processed and remedied by the grinding module.

[0025] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention can quickly distinguish the reflective area in the image by calculating the standard deviation of the color values ​​of the pixels in each pixel block and comparing it with the threshold set by the system, thereby avoiding the influence of the reflective area on the defect detection, and further improving the accuracy of quality detection; by calculating and adjusting the color value of the pixel according to the gray value difference between adjacent pixels, the color value of the pixel around the scratch can be increased and can be highlighted, while also avoiding the edge pixels with too large gray value variation range interfering with the positioning of the scratch, making the scratch detection more accurate; by identifying the approximate edge of the metal workpiece, comparing the color values ​​of the pixels near the approximate edge, screening out the alternative edge pixels, and comparing the detailed parameters of the metal workpiece, the pixels that do not meet the conditions can be eliminated, avoiding Avoid pixels that do not meet the requirements from interfering with the accuracy of edge detection, thereby improving the accuracy of burr detection. By judging the distance between the edge pixel in the metal workpiece image and the feature point of the metal workpiece, it is possible to quickly determine whether the current edge pixel belongs to the metal workpiece, thereby avoiding the situation where there are many and irregular burrs on the metal workpiece, which causes the model to be unable to overlap, resulting in large burr detection errors, greatly improving the accuracy of system detection. By verifying whether there are reflective marks in the scratch pixels, it is possible to avoid the system from misjudging the reflection of metal materials as scratches, further improving the accuracy of system detection. By detecting the width of the workpiece burr and the connection between the workpiece, it is possible to reduce the phenomenon of inaccurate manual judgment that the workpiece can be remedied but is not sent back for remediation, and it is possible to quickly determine whether the current workpiece can be remedied, thereby improving the detection efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0027] Figure 1 It is a schematic diagram of the system module composition of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] See also Figure 1 The present invention provides a technical solution: a quality inspection system based on computer vision, including an image acquisition module, a quality inspection module and a defect processing module, the image acquisition module is used to collect the surface visual image of the processed metal workpiece, collect the metal workpiece detection data and enter various parameter standards of the metal workpiece into the system, the quality inspection module is used to detect the edge burrs, surface scratches and internal cracks of the metal workpiece, the defect processing module is used to process the detected edge burrs and polish the processed edges, and the image acquisition module, the quality inspection module and the defect processing module are electrically connected to each other;

[0030] The image processing module includes a reflection processing submodule and an image normalization processing submodule. The reflection processing submodule is used to process the reflection part of the photographed metal workpiece, and the image normalization processing submodule is used to perform image normalization processing on the collected visual image of the metal workpiece;

[0031] The defect detection module includes a burr detection submodule, a scratch detection submodule and a crack detection submodule. The burr detection submodule is used to detect whether there are burrs on the edge of the metal workpiece, the scratch detection submodule is used to detect whether there are scratches on the metal workpiece, and the crack detection submodule is used to detect whether there are cracks inside the metal workpiece.

[0032] The image acquisition module includes a vision module, a sensor module and a product information entry module. The vision module is used to capture the visual image of the metal workpiece during the processing process, the sensor module is used to measure the edge size of the processed metal workpiece, and the product information entry module is used to enter the comprehensive parameter data of the product produced by the current stamping machine into the system.

[0033] The quality detection module includes an image processing module, which is used to identify shadows and reflections in the image and process the shadow areas and reflection areas in the image to reduce the impact of shadows and reflections on quality detection.

[0034] The defect processing module includes a burr processing module, an edge analysis module and a polishing module. The burr processing module is used to analyze the size, thickness and connection area of ​​the identified burrs, and process the edge burrs of the metal workpiece according to the analysis results. The edge analysis module is used to analyze the cross-section after cutting and analyze the residue of the incision after cutting the burrs. The polishing module is used to polish the workpiece cross-section according to the comprehensive parameters of the metal workpiece.

[0035] The quality inspection module also includes a defect inspection module, which is used to perform quality inspection on the processed metal workpiece.

[0036] The operation method of the quality inspection system mainly includes the following steps:

[0037] Step S1: Using the visual module, a visual image of the stamped metal workpiece on the conveyor belt is captured in real time and uploaded to the system; using the sensor module, the comprehensive detection data collected by the sensor is uploaded to the system; and using the product information entry module, the detailed parameter data of the metal workpiece is entered into the system;

[0038] Step S2: After the data collection is completed, the system starts the image processing module, starts to detect whether there is reflection in the image, and processes the reflection part in the image;

[0039] Step S3: When detecting whether the metal workpiece has defects, the system sends an electrical signal to start the defect detection module, starts analyzing the defects of the metal workpiece, and marks the metal workpiece according to the analysis results;

[0040] Step S4: When processing defects on a metal workpiece, start the burr processing module, start analyzing various parameters of the burr, analyze the burr processing method, and process the burr according to the burr analysis results.

[0041] Step S2 further comprises the following steps:

[0042] Step S21: Retrieving a visual image of the metal workpiece, scanning and identifying the contour nodes of the punching table, cropping the visual image according to the identified contour nodes of the punching table, and normalizing the cropped visual image using a weighted median filtering method;

[0043] Step S22: Segment the processed image in equal proportions, identify the image features in each image block, including color features and brightness features, and calculate the standard deviation of the color values ​​of the pixels in the current image block using the formula Where i = 1, 2, 3...n, F represents the standard deviation of the pixel color value in the current image block, X i Represents the color value of each pixel. It represents the average color value of the metal workpiece surface set by the system. If the standard deviation F of the color value of the pixel in the current image block is greater than the threshold set by the system, the current image block is marked as a reflective area image block, otherwise the current image block is marked as a non-reflective area image block. The mark in the image block is read, the image blocks with the same mark are merged, and the reflective area is identified and marked. By calculating the standard deviation of the color value of the pixel in each pixel block and comparing it with the threshold set by the system, the reflective area in the image can be quickly identified, thereby avoiding the influence of the reflective area on defect detection, and further improving the accuracy of quality inspection.

[0044] Step S3 further comprises the following steps:

[0045] Step S31: retrieve the visual image of the metal workpiece, grayscale the metal workpiece image, identify the grayscale values ​​of all pixels in the current metal workpiece image, record and save it in a collection The value of the enhanced pixel brightness corresponding to the gray value change in the X-axis direction is calculated by the formula S = 2G mn -G m(n+1) , the value of the enhanced pixel brightness corresponding to the change in grayscale value in the Y-axis direction is calculated by the formula W = 2G mn -G (m-1)n If the value S or W of the enhanced pixel brightness corresponding to the gray value change is greater than the maximum threshold set by the system, the current pixel is marked as a scratch pixel, otherwise the current pixel is removed. By calculating and adjusting the gray value of the pixel according to the gray value difference between adjacent pixels, the gray value of the pixels around the scratch can be increased and highlighted. At the same time, it also prevents the pixels at the edge of the workpiece with a large gray value change range from interfering with the positioning of the scratch, making the scratch detection more accurate.

[0046] Step S32: Identify the edge feature nodes of the metal workpiece image, mark the first edge of the current metal workpiece according to the edge feature nodes of the metal workpiece, identify the color value of the first edge pixel of the metal workpiece, and when the difference between the color value of the edge pixel of the metal workpiece and the color value of the adjacent pixel is greater than the threshold set by the system, mark the current pixel as an edge pixel, retrieve the color value of the edge pixel, and compare the color value of the edge pixel with the threshold set by the system. If the color value of the current edge pixel is within the threshold range set by the system, retain the mark of the current edge pixel, otherwise delete the edge pixel mark of the current pixel, retrieve the marked edge pixel, connect the adjacent edge pixels, and construct an edge model of the metal workpiece. By identifying the approximate edge of the metal workpiece, compare the color values ​​of the pixels near the approximate edge, screen out the candidate edge pixels, and compare the detailed parameters of the metal workpiece, the pixels that do not meet the conditions can be eliminated to avoid the interference of the pixels that do not meet the requirements on the accuracy of edge detection, thereby improving the accuracy of burr detection;

[0047] Step S33: retrieve the visual image of the metal workpiece, scan and anchor the feature points on the metal workpiece, establish a coordinate system in the image, and calculate the distance between the edge pixel point in the current metal workpiece image and the feature point of the metal workpiece by using a formula Wherein, L represents the distance between the current edge pixel point and the feature point of the metal workpiece, (X1, Y1) represents the coordinates of the current edge pixel point, (X2, Y2) represents the coordinates of the feature node of the metal workpiece, if the distance L between the edge pixel point in the current metal workpiece image and the feature point of the metal workpiece does not match the threshold set by the system, the current edge pixel point is marked as a burr, otherwise the system continues to detect, identifies the edge pixel point marked with the burr, and calculates the distance L2 between the adjacent edge pixel points marked with the burr, if L2 is less than the threshold set by the system, the two adjacent edge pixels are connected to construct a burr model, otherwise the system continues to detect, by judging the distance between the edge pixel point in the metal workpiece image and the feature point of the metal workpiece, it is possible to quickly determine whether the current edge pixel point belongs to the metal workpiece, avoid the metal workpiece burrs are many and irregular, resulting in the model cannot overlap, resulting in large burr detection errors, greatly improving the accuracy of system detection.

[0048] In step S31, the marked scratch pixel points are retrieved to identify whether the scratch pixel points are reflective pixels. If the current scratch pixel points have reflective marks, the scratch marks in the current scratch pixel points are deleted. Otherwise, the scratch pixels are connected to construct a scratch model, and the color value of the pixel with the largest pixel color value in the scratch model is measured. If the color value of this pixel point is less than the system-set threshold, there may be a crack at the location of the current pixel point, and the current pixel point color value is marked. Otherwise, the system continues to detect and retrieve ultrasonic detection data, and compares the ultrasonic detection data with the database. If the similarity is less than the threshold, it is marked that the current metal material has cracks, otherwise there are no cracks. By verifying whether there are reflective marks in the scratch pixels, the system can avoid misjudging the reflection of metal materials as scratches, thereby further improving the accuracy of system detection.

[0049] In step S4, the burr detection data of the metal workpiece is retrieved, the anchor is located at the burr edge node on the edge of the metal workpiece, the edge turning angle of the metal workpiece between the two current burr edge nodes is identified, and the length L3 that needs to be cut for the current burr is calculated according to the formula. If the current cutting length L3 is greater than the threshold set by the system, the current metal workpiece is marked as scrapped, otherwise the current metal workpiece is marked as a remediable workpiece, and the remediable workpiece is retrieved and processed and remedied by the grinding module. By detecting the width of the workpiece burr and the connection between the workpiece, the phenomenon of inaccurate manual judgment that the workpiece can be remedied but is not sent back for remediation can be reduced, and it can quickly determine whether the current workpiece can be remedied, thereby improving the detection efficiency of the system.

[0050] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0051] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A quality inspection system based on computer vision, including an image acquisition module, a quality inspection module and a defect processing module, characterized in that: The image acquisition module is used to collect the surface visual image of the processed metal workpiece, collect the metal workpiece detection data and enter various parameter standards of the metal workpiece into the system; the quality inspection module is used to detect the edge burrs, surface scratches and internal cracks of the metal workpiece; the defect processing module is used to process the detected edge burrs and polish the processed edges; the image acquisition module, the quality inspection module and the defect processing module are electrically connected to each other; The quality detection module includes an image processing module, which is used to identify shadows and reflections in the image and process the shadow areas and reflection areas in the image to reduce the impact of the shadows and reflections on the quality detection; The quality inspection module also includes a defect inspection module, which is used to perform quality inspection on the processed metal workpiece; The image processing module includes a reflection processing submodule and an image normalization processing submodule, wherein the reflection processing submodule is used to process the reflection part of the photographed metal workpiece, and the image normalization processing submodule is used to perform image normalization processing on the collected visual image of the metal workpiece; The defect detection module includes a burr detection submodule, a scratch detection submodule and a crack detection submodule. The burr detection submodule is used to detect whether there are burrs on the edge of the metal workpiece, the scratch detection submodule is used to detect whether there are scratches on the metal workpiece, and the crack detection submodule is used to detect whether there are cracks inside the metal workpiece. The image acquisition module includes a visual module, a sensor module and a product information entry module, wherein the visual module is used to capture the visual image of the metal workpiece during processing, the sensor module is used to measure the edge size of the processed metal workpiece, and the product information entry module is used to enter the comprehensive parameter data of the product produced by the current stamping machine into the system; The defect processing module includes a burr processing module, an edge analysis module and a grinding module. The burr processing module is used to analyze the size, thickness and connection area of ​​the identified burrs, and process the edge burrs of the metal workpiece according to the analysis results. The edge analysis module is used to analyze the cross section after cutting and analyze the residue of the cut after cutting the burrs. The grinding module is used to grind the workpiece cross section according to the comprehensive parameters of the metal workpiece; The operation method of the quality detection system mainly includes the following steps: Step S1: Using the visual module, a visual image of the stamped metal workpiece on the conveyor belt is captured in real time and uploaded to the system; using the sensor module, the comprehensive detection data collected by the sensor is uploaded to the system; and using the product information entry module, the detailed parameter data of the metal workpiece is entered into the system; Step S2: After the data collection is completed, the system starts the image processing module, starts to detect whether there is reflection in the image, and processes the reflection part in the image; Step S3: When detecting whether the metal workpiece has defects, the system sends an electrical signal to start the defect detection module, starts analyzing the defects of the metal workpiece, and marks the metal workpiece according to the analysis results; Step S4: when processing defects on a metal workpiece, start the burr processing module, start analyzing various parameters of the burr, analyze the burr processing method, and process the burr according to the burr analysis results; The step S2 further comprises the following steps: Step S21: Retrieving a visual image of the metal workpiece, scanning and identifying the contour nodes of the punching table, cropping the visual image according to the identified contour nodes of the punching table, and normalizing the cropped visual image using a weighted median filtering method; Step S22: Segment the processed image in equal proportions, identify the image features in each image block, including color features and brightness features, and calculate the standard deviation of the color values ​​of the pixels in the current image block using the formula Where i = 1, 2, 3...n, F represents the standard deviation of the pixel color value in the current image block, X i Represents the color value of each pixel. Indicates the average value of the color value of the metal workpiece surface set by the system. If the standard deviation F of the color value of the pixel point in the current image block is greater than the threshold set by the system, the current image block is marked as a reflective area image block, otherwise the current image block is marked as a non-reflective area image block. The mark in the image block is read, and the image blocks with the same mark are merged to identify and mark the reflective area. The step S3 further comprises the following steps: Step S31: retrieve the visual image of the metal workpiece, grayscale the metal workpiece image, identify the grayscale values ​​of all pixels in the current metal workpiece image, record and save it in a collection The value of the enhanced pixel brightness corresponding to the gray value change in the X-axis direction is calculated by the formula S = 2G mn -G m(n+1) , the value of the enhanced pixel brightness corresponding to the change in grayscale value in the Y-axis direction is calculated by the formula W = 2G mn -G (m-1)n , if the value S or W of the enhanced pixel brightness corresponding to the gray value change is greater than the maximum threshold set by the system, the current pixel is marked as a scratch pixel, otherwise the current pixel is removed; Step S32: Identify edge feature nodes of the metal workpiece image, mark the first edge of the current metal workpiece according to the edge feature nodes of the metal workpiece, identify the color value of the first edge pixel of the metal workpiece, and when the difference between the color value of the edge pixel of the metal workpiece and the color value of the adjacent pixel is greater than the threshold set by the system, mark the current pixel as an edge pixel, retrieve the color value of the edge pixel, and compare the color value of the edge pixel with the threshold set by the system. If the color value of the current edge pixel is within the threshold range set by the system, retain the mark of the current edge pixel, otherwise delete the edge pixel mark of the current pixel, retrieve the marked edge pixel, connect the adjacent edge pixels, and construct an edge model of the metal workpiece; Step S33: retrieve the visual image of the metal workpiece, scan and anchor the feature points on the metal workpiece, establish a coordinate system in the image, and calculate the distance between the edge pixel point in the current metal workpiece image and the feature point of the metal workpiece by using a formula Wherein, L represents the distance between the current edge pixel point and the feature point of the metal workpiece, (X1, Y1) represents the coordinates of the current edge pixel point, (X2, Y2) represents the coordinates of the feature node of the metal workpiece, if the distance L between the edge pixel point in the current metal workpiece image and the feature point of the metal workpiece does not match the threshold set by the system, the current edge pixel point is marked as a burr, otherwise the system continues to detect, identify the edge pixel point marked with the burr, and calculate the distance L2 between the adjacent edge pixel points marked with the burr, if L2 is less than the threshold set by the system, the two adjacent edge pixels are connected to construct a burr model, otherwise the system continues to detect; In the step S31, the marked scratch pixel points are retrieved to identify whether the scratch pixel points are reflective pixel points. If the current scratch pixel points have reflective marks, the scratch marks in the current scratch pixel points are deleted. Otherwise, the scratch pixels are connected to build a scratch model, and the color value of the pixel point with the largest pixel color value in the scratch model is measured. If the color value of this pixel point is less than the threshold set by the system, the color value of the current pixel point is marked. Otherwise, the system continues to detect and retrieve ultrasonic detection data, and compares the ultrasonic detection data with the database. If the similarity is less than the threshold, it is marked that the current metal material has cracks, otherwise, no cracks exist; In the step S4, the burr detection data of the metal workpiece is retrieved, the anchor is located at the burr edge node on the edge of the metal workpiece, the edge turning angle of the metal workpiece between the two current burr edge nodes is identified, and the length L3 that needs to be cut of the current burr is calculated according to the formula. If the current cutting length L3 is greater than the threshold set by the system, the current metal workpiece is marked as scrapped, otherwise the current metal workpiece is marked as a remediable workpiece, and the remediable workpiece is retrieved and processed and remediated by the grinding module.

Images