Intelligent identification method for automobile connecting plate processing

Abstract

The invention belongs to the technical field of image processing, and particularly relates to an intelligent recognition method for automobile connecting plate processing, which comprises the following steps of: acquiring a surface image of a corner position of a processed automobile connecting plate, and performing threshold segmentation on the surface image through a target low threshold and a target high threshold respectively to obtain a target threshold; the method comprises the steps of obtaining a low binary image, determining a guide line contour from the obtained low binary image, calculating a first probability that a connecting plate has a wrinkling defect according to the curvature change condition of contour points on the guide line contour, determining an intersection angle contour from the obtained high binary image, and calculating a second probability that the connecting plate has a wrinkling defect according to the difference of chain code sets of contour branch points on the intersection angle contour in different directions. And calculating the second probability that the connecting plate has the wrinkling defect, and judging whether the automobile connecting plate has the wrinkling defect or not according to the first probability and the second probability that the connecting plate has the wrinkling defect. According to the invention, the wrinkling defect can be detected more accurately, and the product quality and the production efficiency of the automobile connecting plate are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing. More specifically, the present invention relates to an intelligent recognition method for processing automotive connection plates. Background Art

[0002] Automotive floor crossbeam connection plates are usually installed at the front, middle, and rear of the vehicle body chassis. It is connected to other structural components, such as floor panels, vehicle body longitudinal beams, and floor crossbeams, etc., to jointly form the vehicle chassis structure and is a key connecting component of the vehicle body skeleton.

[0003] This connection plate is mainly made of high-strength steel. The flange surface at the bottom, the vertical surface at the end, and the connection surface at the top are respectively connected to the floor, the side wall, and the top surface of the vehicle body longitudinal beam through welding points.

[0004] In order to improve the load-bearing capacity, impact resistance, and torsional rigidity, and thus enhance the safety of the vehicle body and the stability of the chassis, the product structure design of the floor crossbeam connection plate increasingly tends to adopt closed connection surfaces to prevent tearing at the notch during a collision. However, this design also increases the difficulty of stamping and forming.

[0005] When producing the floor crossbeam connection plate using a forming process, the blank at the three-sided intersection position is extruded and deformed, and bulges at the corner position. After the die is closed, the excess material cannot be eliminated, resulting in a wrinkling defect at the wrinkled area at the corner position, which in turn affects When producing the floor crossbeam connection plate using a forming process, the blank undergoes extrusion deformation at the three-sided intersection position, causing bulging at the corner; after the die is closed, the excess material cannot be effectively removed, resulting in a wrinkling defect at the corner position.

[0006] The wrinkling defect will weaken the structural strength and rigidity of the connection plate, reduce its fatigue resistance, and thus affect the overall stability and safety of the vehicle body; at the same time, the wrinkling problem will also cause the quality of the produced connection plate to be unstable, increase the defective rate, reduce the production efficiency, and bring additional costs and quality control challenges to the manufacturing process. Summary of the Invention

[0007] To solve the above technical problem of how to determine whether there is a wrinkling defect in the processed automotive connecting plate, the present invention provides an intelligent recognition method for processing automotive connecting plates, including: collecting the surface image of the corner position of the processed automotive connecting plate; updating the initial low threshold multiple times, and determining the target low threshold according to the change in the number and connection of the black connected regions in the binary images corresponding to the low thresholds before and after each update; performing threshold segmentation on the surface image through the target low threshold, extracting the skeleton of the largest black connected region in the obtained low binary image to obtain the guide line contour; calculating the first probability of the connecting plate having a wrinkling defect according to the change in the curvature of the contour points on the guide line contour; updating the initial high threshold multiple times, and determining the target high threshold according to the change in the number and connection of the white connected regions in the binary images corresponding to the high thresholds before and after each update; performing threshold segmentation on the surface image through the target high threshold, extracting the skeleton of the white connected region in the obtained high binary image to obtain the intersection angle contour of the surface image; obtaining the contour branch points on the intersection angle contour that contain multiple neighborhood points in the 8-neighborhood, extracting the chain code of the intersection angle contour based on the contour branch points, and calculating the second probability of the connecting plate having a wrinkling defect according to the difference in the chain code sets of the contour branch points in different directions; judging whether the processed automotive connecting plate has a wrinkling defect according to the first probability and the second probability.

[0008] Preferably, the step of updating the initial low threshold multiple times and determining the target low threshold according to the change in the number and connection of the black connected regions in the binary images corresponding to the low thresholds before and after each update includes: updating the initial low threshold multiple times, performing threshold segmentation on the surface image according to the low threshold after each update to obtain the binary image corresponding to the low threshold after each update; performing connected region analysis on all black pixel points in the binary image corresponding to the low threshold after each update to obtain multiple black connected regions; until the two largest black connected regions in the binary image corresponding to the low threshold after the -th update are merged into one black connected region in the binary image corresponding to the low threshold after the -th update, stop updating the initial low threshold, and use the low threshold after the -th update as the target low threshold.

[0009] Preferably, the method for obtaining the curvature at the contour point includes: for any contour point on the guide line contour, linearly fitting the positions of the contour point and its 6 adjacent points, and using the slope of the obtained straight line as the curvature at the contour point.

[0010] Preferably, the step of calculating the first probability of the connecting plate having a wrinkling defect according to the change in the curvature of the contour points on the guide line contour includes: ; where is the first probability of the connecting plate having a wrinkling defect, is the natural exponential function, is the curvature of the th contour point on the guide line contour, is the curvature of the th contour point on the guide line contour, is the number of all contour points included in the guide line contour.

[0011] Preferably, the initial high threshold is updated multiple times, and according to the change in the number and connection of white connected regions in the binary image corresponding to the high threshold before and after each update, the target high threshold is determined, including: updating the initial high threshold multiple times, performing threshold segmentation on the surface image according to the high threshold after each update, and obtaining the binary image corresponding to the high threshold after each update; performing connected region analysis on all white pixel points in the binary image corresponding to the high threshold after each update to obtain multiple white connected regions; until there is only 1 white connected region in the binary image corresponding to the high threshold after the th update , and this white connected region is segmented into at least 2 white connected regions in the binary image corresponding to the high threshold after the th update, stop updating the initial high threshold, and use the high threshold after the th update as the target high threshold.

[0012] Preferably, the chain code extraction of the intersection angle contour based on the contour branch point includes: for any contour branch point, along the direction corresponding to any neighborhood point within the 8-neighborhood of this contour branch point, by judging whether there are other contour points within the 8-neighborhood of each contour point, if there are other contour points within the 8-neighborhood, obtain the chain code value of the other contour point compared to this contour point, and sequentially obtain the chain code values of each contour point until there are no other contour points within the 8-neighborhood of the contour point or the other contour points within the 8-neighborhood of the contour point belong to the contour branch point, and form the chain code set of this contour branch point in this direction with all the chain code values.

[0013] Preferably, calculating the second probability of the connecting plate having a wrinkling defect according to the difference between the chain code sets of the contour branch point in different directions includes: ; where is the second probability of the connecting plate having a wrinkling defect, is the number of all contour branch points on the intersection angle contour, is the maximum value of the angles between the standard chain code values of the th contour branch point in every two different directions, is the natural exponential function.

[0014] Preferably, the method for obtaining the standard chain code values of the contour branch points in all directions includes: for any one contour branch point, calculating the mean value of all the chain code values in the chain code set in all directions of this contour branch point as the standard chain code value of this contour branch point in all directions.

[0015] Preferably, judging whether there is a wrinkling defect in the processed automotive connecting plate according to the first probability and the second probability includes: when the first probability that the connecting plate has a wrinkling defect and the second probability that the connecting plate has a wrinkling defect the maximum value of is greater than the tolerance

[0016] it is considered that the connecting plate has a wrinkling defect; otherwise, according to the first probability and the second probability that the connecting plate has a wrinkling defect, calculate the comprehensive probability that the connecting plate has a wrinkling defect. When the comprehensive probability that the connecting plate has a wrinkling defect is greater than the tolerance it is considered that the connecting plate has a wrinkling defect. Preferably, the calculation formula for the comprehensive probability that the connecting plate has a wrinkling defect is: ; where is the comprehensive probability that the connecting plate has a wrinkling defect,

[0017] The beneficial effects of the present invention are as follows: Based on the two characteristics that the wrinkling defect will affect the guiding lines corresponding to other assembly relationships near the corner position of the connecting plate and will generate new protrusions near the corner position of the connecting plate, the present invention respectively performs threshold segmentation on the surface image through the target low threshold and the target high threshold, determines the guiding line contour from the obtained low binary image, calculates the first probability that the connecting plate has a wrinkling defect according to the curvature change of the contour points on the guiding line contour, determines the intersection angle contour from the obtained high binary image, and calculates the second probability that the connecting plate has a wrinkling defect according to the difference of the chain code sets of each contour branch point on the intersection angle contour in different directions; and then judges whether there is a wrinkling defect in the processed automotive connecting plate according to the first probability and the second probability that the connecting plate has a wrinkling defect, can detect the wrinkling defect more accurately, improve the product quality and production efficiency of the automotive connecting plate, reduce the defective rate and production cost, and is applicable to the on-line detection and quality control of automotive connecting plates. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 ​It is a flowchart schematically showing an intelligent recognition method for processing an automotive connecting plate in the present invention; Figure 2 It is a schematic diagram schematically showing the corner position at the three-sided intersection angle of the connecting plate; Figure 3 It is a schematic diagram of the surface image of the corner position of a normal automotive connecting plate; Figure 4 It is a schematic diagram of the surface image of the corner position of an automotive connecting plate with a wrinkling defect; Figure 5 It is a flowchart schematically showing step S2; Figure 6 It is schematically showing Figure 3 A schematic diagram of the black connected domain in the binary image corresponding to the low threshold after different updates of the shown surface image; Figure 7 It is schematically showing Figure 4 A schematic diagram of the binary image corresponding to the low threshold after different updates of the shown surface image; Figure 8 It is schematically showing Figure 3 A schematic diagram of the binary image corresponding to the high threshold after different updates of the shown surface image; Figure 9 It is schematically showing Figure 4 A schematic diagram of the binary image corresponding to the high threshold after different updates of the shown surface image. Specific embodiments

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of protection of the present invention.

[0020] The specific embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0021] An embodiment of the present invention discloses an intelligent recognition method for processing an automotive connecting plate. Referring to Figure 1 , it includes steps S1 to S3: S1. Collect the surface image of the corner position of the processed automotive connecting plate.

[0022] It should be noted that the forming process plan is adopted for the production of the floor beam connection plate. The material is pressed in the middle area and formed freely on both sides. Since the two side walls with an included angle of about 90 degrees are formed downward at the same time, and the blank holder core at the R corner position of the product is in a closed state, the blank at the three-sided intersection angle is extruded and deformed, and bulges at the corner position. After the mold is closed, the excess material cannot be eliminated, resulting in a wrinkling defect at the wrinkled part at the corner position. Therefore, in order to detect whether there is a wrinkling defect in the processed automotive connection plate, it is necessary to collect the surface image of the corner position of the connection plate, and judge whether there is a wrinkling defect in the automotive connection plate by analyzing the surface image.

[0023] Specifically, on the production line for manufacturing the automotive floor beam connection plate, after the stamping part is formed and separated from the mold, a camera is used to capture the surface image of the corner position at the three-sided intersection of the processed automotive connection plate. Among them, the schematic diagram of the corner position at the three-sided intersection of the connection plate is as Figure 2 shown.

[0024] S2. Respectively, through the target low threshold and the target high threshold, perform threshold segmentation on the surface image, determine the guide line contour from the obtained low binary image, calculate the first probability of the connection plate having a wrinkling defect according to the curvature change of the contour points on the guide line contour, determine the intersection angle contour from the obtained high binary image, and calculate the second probability of the connection plate having a wrinkling defect according to the difference in the chain code sets of each contour branch point on the intersection angle contour in different directions.

[0025] It should be noted that the wrinkling defect has two aspects of influence on the corner position of the connection plate: (1) On the one hand, the wrinkling defect will affect the guide line corresponding to other assembly relationships near the corner position of the connection plate, resulting in the guide line becoming more compact, and finally manifested as a smaller angle at the turning point of the guide line. As Figure 3 shown, it is a schematic diagram of the surface image of the corner position of a normal automotive connection plate. Among them, the area marked by the white frame is the guide line corresponding to other assembly relationships near the corner position of the connection plate; as Figure 4 shown, it is a schematic diagram of the surface image of the corner position of an automotive connection plate with a wrinkling defect. Among them, the area marked by the white frame is the guide line affected by the wrinkling defect.

[0026] (2) On the other hand, the wrinkling defect will generate new protrusions near the corner position of the connection plate and connect with the contour of the three-sided intersection angle at the corner position, and finally manifested as an increase in the branches on the contour formed by brighter pixel points. As Figure 3 shown, it is a schematic diagram of the surface image of the corner position of a normal automotive connection plate. Among them, due to the existence of two side walls with an included angle of about 90°, and the existence of a bottom wall with an included angle of about 90° with the two side walls, the angle of the branches on the contour formed by brighter pixel points is close to 90°; asFigure 4 As shown in the figure, it is a schematic diagram of the surface image of the corner position of an automotive connecting plate with wrinkling defects. The wrinkling defects will cause an increase in the branches on the contour formed by brighter pixel points, and the angles of the increased branches do not meet the condition of being close to 90°.

[0027] In summary, by the angle at the turning point of the guiding line, the change in the number of branches on the contour formed by brighter pixel points, and the change in the angle, it is possible to determine whether wrinkling defects have occurred at the corner position of the connecting plate.

[0028] The flowchart of step S2 is referred to Figure 5 , including step S201 to step S202, specifically: S201. Perform threshold segmentation on the surface image through the target low threshold, determine the guiding line contour from the obtained low binary image, and calculate the first probability of the connecting plate having wrinkling defects according to the curvature change of the contour points on the guiding line contour.

[0029] 1. Update the initial low threshold multiple times, and determine the target low threshold according to the change in the number and connection of the black connected regions in the binary images corresponding to the low thresholds before and after each update.

[0030] Specifically, update the initial low threshold multiple times, perform threshold segmentation on the surface image according to the low threshold after each update, and obtain the binary image corresponding to the low threshold after each update; perform connected component analysis on all the black pixel points in the binary image corresponding to the low threshold after each update to obtain multiple black connected regions; denote the two black connected regions with the largest areas as and ; until in the binary image corresponding to the low threshold after the -th update, the two black connected regions and , in the binary image corresponding to the low threshold after the -th update, are merged into 1 black connected region, stop updating the initial low threshold, and use the low threshold after the -th update as the target low threshold.

[0031] Among them, the specific value of the initial low threshold can be set according to the actual application scenario and requirements, and the value range of the initial low threshold is [85, 93]. In the present invention, the initial low threshold is set to 90.

[0032] Exemplarily, for the surface image of the corner position of a normal automotive connecting plate as shown in Figure 3 , in the binary image corresponding to the low threshold after the 4th update, the two black connected regions with the largest areas obtained are as shown in Figure 6As shown in (1), the two black connected regions are merged into one black connected region in the binary image corresponding to the low threshold after the 5th update, as Figure 6 shown in (2). Therefore, the low threshold after the 5th update is used as the target low threshold.

[0033] Exemplarily, for Figure 4 the surface image of the corner position of the automotive connecting plate with wrinkling defects shown, in the binary image corresponding to the low threshold after the 5th update, the two black connected regions with the largest areas obtained are as Figure 7 shown in (1). The two black connected regions are merged into one black connected region in the binary image corresponding to the low threshold after the 6th update, as Figure 7 shown in (2). Therefore, the low threshold after the 6th update is used as the target low threshold.

[0034] 2. Perform threshold segmentation on the surface image using the target low threshold to obtain a low binary image.

[0035] Specifically, the pixel points with gray values less than or equal to the target low threshold are marked as black pixel points, and the pixel points with gray values greater than the target low threshold are marked as white pixel points, thereby obtaining a low binary image.

[0036] Exemplarily, for Figure 3 the surface image of the corner position of the normal automotive connecting plate shown, the schematic diagram of the low binary image obtained by performing threshold segmentation on the surface image using the target low threshold is as Figure 6 shown in (2). For Figure 4 the surface image of the corner position of the automotive connecting plate with wrinkling defects shown, the schematic diagram of the low binary image obtained by performing threshold segmentation on the surface image using the target low threshold is as Figure 7 shown in (2).

[0037] 3. Perform skeleton extraction on the black connected region with the largest area in the low binary image to obtain the guiding line contour of the surface image.

[0038] Among them, skeleton extraction is a technique in image processing used to extract the center line or contour of an object from a binary image; since wrinkling defects will cause the angle at the turning point of the guiding line to become smaller, therefore, it is necessary to perform skeleton extraction on the black connected region with the largest area in the low binary image to obtain the guiding line contour of the surface image.

[0039] The methods for performing skeleton extraction include, but are not limited to: the K3M algorithm (Kirschner-Mosteller-Kalaba), the Medial Axis Transform (MAT), the Medial Axis Transform (MAT), the Distance Transform with Thinning, the Topological Skeleton Extraction, and the Fast Marching Method (FMM).

[0040] 4. Calculate the first probability of the connecting plate having a wrinkling defect according to the curvature change of the contour points on the guiding line contour.

[0041] Specifically, for any contour point on the guiding line contour, linearly fit the positions of this contour point and its six adjacent points, and use the slope of the obtained straight line as the curvature at this contour point.

[0042] Furthermore, calculate the first probability of the connecting plate having a wrinkling defect according to the curvature change of the contour points on the guiding line contour. The calculation formula for the first probability of the connecting plate having a wrinkling defect is as follows: ; In the formula, is the first probability of the connecting plate having a wrinkling defect, is the natural exponential function, is the curvature of the th contour point on the guiding line contour, is the curvature of the th contour point on the guiding line contour. The th contour point and the th contour point on the guiding line contour are adjacent contour points, is the number of all contour points included in the guiding line contour, represents the maximum value function, represents the arctangent function, which is used to determine the included angle through the curvature.

[0043] It should be noted that in the surface image of the automotive connecting plate, the angle of the contour point representing the turning point of the guiding line will change suddenly. Therefore, by calculating the difference in the included angle between two adjacent contour points, that is , and by obtaining the maximum value among the differences in the included angles between each two adjacent contour points, that is , obtain contour points representing the turning points of the guiding line from the guiding line contour; on this basis, since the wrinkling defect will cause the guiding line to become more compact, and finally manifested as a smaller angle at the turning point of the guiding line, therefore, the smaller the included angle of the contour points representing the turning point of the guiding line in the surface image of the automotive connecting plate, the greater the probability that the connecting plate has a wrinkling defect, that is, the first probability that the connecting plate has a wrinkling defect The larger; in addition, the used for to perform normalization.

[0044] S202. Perform threshold segmentation on the surface image through the target high threshold, determine the intersection angle contour from the obtained high binary image, and calculate the second probability that the connecting plate has a wrinkling defect according to the difference in the chain code sets of each contour branch point on the intersection angle contour in different directions.

[0045] 1. Update the initial high threshold multiple times, and determine the target high threshold according to the change in the number of white connected regions and the connection change in the binary image corresponding to the high threshold before and after each update.

[0046] Specifically, update the initial high threshold multiple times, perform threshold segmentation on the surface image according to the high threshold after each update, and obtain the binary image corresponding to the high threshold after each update; perform connected region analysis on all white pixel points in the binary image corresponding to the high threshold after each update to obtain multiple white connected regions; until there is only 1 white connected region in the binary image corresponding to the high threshold after the th update and this white connected region in the th update of the binary image corresponding to the high threshold is divided into at least 2 white connected regions, stop updating the initial high threshold, and use the high threshold after the th update as the target high threshold.

[0047] Among them, the specific value of the initial high threshold can be set according to the actual application scenario and requirements, and the value range of the initial high threshold is [95, 115]. In the present invention, the initial high threshold is set to 110.

[0048] Exemplarily, for the surface image of the corner position of the normal automotive connecting plate shown in Figure 3 , in the binary image corresponding to the high threshold after the 6th update, there is only 1 white connected region as shown in Figure 8 in (1) therein, and this white connected region is divided into at least 2 white connected regions in the binary image corresponding to the high threshold after the 7th update, as shown in Figure 8 in (1) therein. Therefore, the high threshold after the 6th update is used as the target high threshold.

[0049] Exemplarily, for Figure 3 the surface image at the corner position of the automotive connecting plate with a wrinkling defect as shown, in the binary image corresponding to the high threshold after the 3rd update, there is only 1 white connected region as Figure 9 shown in (1) below. This white connected region is split into at least 2 white connected regions in the binary image corresponding to the high threshold after the 7th update, as Figure 9 shown in (2) below. Therefore, the high threshold after the 3rd update is taken as the target high threshold.

[0050] 2. Perform threshold segmentation on the surface image using the target high threshold to obtain a high binary image.

[0051] Specifically, the pixel points with gray values less than or equal to the target high threshold are marked as black pixel points, and the pixel points with gray values greater than the target high threshold are marked as white pixel points, thereby obtaining a high binary image, and there is exactly one white connected region in the obtained high binary image.

[0052] Exemplarily, for Figure 3 the surface image at the corner position of a normal automotive connecting plate as shown, the schematic diagram of the high binary image obtained by performing threshold segmentation on the surface image using the target high threshold is as Figure 8 shown in (1) below. For Figure 4 the surface image at the corner position of the automotive connecting plate with a wrinkling defect as shown, the schematic diagram of the high binary image obtained by performing threshold segmentation on the surface image using the target high threshold is as Figure 9 shown in (1) below.

[0053] 3. Perform skeleton extraction on the white connected region in the high binary image to obtain the intersection angle contour of the surface image.

[0054] Among them, since the wrinkling defect will generate new protrusions near the corner position of the connecting plate, resulting in an increase in the branches on the contour formed by the brighter pixel points, therefore, it is necessary to perform skeleton extraction on the white connected region in the high binary image to obtain the guiding line contour of the surface image.

[0055] 4. Obtain all the contour branch points on the intersection angle contour, and perform chain code extraction on the intersection angle contour based on the contour branch points to obtain the set of chain codes of the contour branch points in each direction.

[0056] It should be noted that on the automotive connecting plate, the contours of the three-sided intersection angle converge at the corner position, and the included angle between the two side walls is approximately 90°, and the included angles between the bottom wall and the two side walls are approximately 90°. Therefore, first, according to the number of neighborhood points within the 8-neighborhood of the contour points on the intersection angle contour, the contour branch points are obtained to represent the convergence of the contours of the three-sided intersection angle; then, based on the convergence of the contours of the three-sided intersection angle, that is, the contour branch points, chain code extraction is performed on the intersection angle contour to obtain the set of chain codes in each direction of the contour branch points, which is used to represent the multiple branch contours that make up the three-sided intersection angle.

[0057] Specifically, obtain the number of neighborhood points within the 8-neighborhood of each contour point on the intersection angle contour. If the number of neighborhood points within the 8-neighborhood of the contour point is greater than 1, it means that there are multiple neighborhood points within the 8-neighborhood of the contour point. Denote the contour points with multiple neighborhood points within the 8-neighborhood as contour branch points.

[0058] Furthermore, for any one contour branch point, along the direction corresponding to any one neighborhood point within the 8-neighborhood of this contour branch point, by judging whether there are other contour points within the 8-neighborhood of each contour point. If there are other contour points within the 8-neighborhood, obtain the chain code value of the other contour point relative to this contour point, and sequentially obtain the chain code values of each contour point until there are no other contour points within the 8-neighborhood of the contour point or the other contour points within the 8-neighborhood of the contour point belong to the contour branch point and then stop. Combine all the chain code values to form the set of chain codes of this contour branch point in this direction. The set of chain codes of this contour branch point in this direction represents a branch contour that makes up the three-sided intersection angle.

[0059] In this way, obtain the set of chain codes of each contour branch point in the directions corresponding to each neighborhood point within the 8-neighborhood.

[0060] 6. According to the set of chain codes of the contour branch points in each direction, obtain the standard chain code values of the contour branch points in each direction.

[0061] It should be noted that the set of chain codes of the contour branch points in each direction can represent the multiple branch contours that make up the three-sided intersection angle. Then, the mean value of all the chain code values in the set of chain codes of the contour branch points in each direction is used to represent the direction trend of the multiple branch contours converging at the contour branch point. Furthermore, by judging whether the angles between the direction trends of the multiple branch contours satisfy the condition of being close to 90°, it is determined whether the multiple branch contours obtained are indeed the branch contours that make up the three-sided intersection angle, and further determine whether the contour branch point is indeed the convergence of the contours of the three-sided intersection angle.

[0062] It should be further noted that the wrinkling defect will generate new protrusions near the corner position of the connecting plate and connect with the contour of the three-sided intersection angle at the corner position. Finally, it is manifested as an increase in the branches on the contour formed by relatively bright pixel points. Therefore, if there are contour branch points that do not belong to the contour convergence of the three-sided intersection angle, it indicates that the automotive connecting plate has a wrinkling defect.

[0063] Specifically, for any contour branch point, calculate the mean value of all chain code values in the chain code set in each direction of this contour branch point as the standard chain code value of this contour branch point in each direction.

[0064] In this way, obtain the standard chain code values of each contour branch point in the directions corresponding to each neighborhood point within the 8-neighborhood.

[0065] 7. Calculate the second probability of the existence of wrinkling defects in the connecting plate according to the differences in the chain code sets of each contour branch point on the intersection angle contour in different directions.

[0066] The specific calculation formula is: ; In the formula, is the second probability of the existence of wrinkling defects in the connecting plate, is the number of all contour branch points on the intersection angle contour, is the maximum value of the included angle between the standard chain code values of the th contour branch point in every two different directions,

[0067] It should be noted that the closer the maximum value of the included angle between the standard chain code values of the contour branch point in every two different directions is to 90°, the more likely this contour branch point belongs to the contour branch point of the contour convergence of the three-sided intersection angle. At this time is closer to 1. On the contrary, the more the maximum value of the included angle between the standard chain code values of the contour branch point in every two different directions deviates from 90°, the less likely this contour branch point belongs to the contour branch point of the contour convergence of the three-sided intersection angle. Correspondingly, the probability of the existence of wrinkling defects in the connecting plate is greater. At this time is closer to 0, is closer to 1, and correspondingly, the second probability of the existence of wrinkling defects in the connecting plate is

[0068]

[0069] 1. When the maximum value of the first probability of the existence of wrinkling defects in the connecting plate and the second probability of the existence of wrinkling defects in the connecting plate is greater than the tolerance When it is, it is considered that there is a wrinkling defect in the connecting plate.

[0070] Among them, the tolerance The specific value of can be set according to the actual application scenario and requirements, and the tolerance The value range of is [0.85, 1), and the present invention sets the preset first threshold to 0.9.

[0071] 2. Otherwise, according to the first probability and the second probability that the connecting plate has a wrinkling defect, calculate the comprehensive probability that the connecting plate has a wrinkling defect. When the comprehensive probability that the connecting plate has a wrinkling defect is greater than the tolerance When it is, it is considered that there is a wrinkling defect in the connecting plate.

[0072] Among them, the calculation formula for the comprehensive probability that the connecting plate has a wrinkling defect is: ; In the formula, is the comprehensive probability that the connecting plate has a wrinkling defect, represents the maximum value function, represents the minimum value function; is the first probability that the connecting plate has a wrinkling defect, is the second probability that the connecting plate has a wrinkling defect.

[0073] It should be noted that according to the first probability and the second probability that the connecting plate has a wrinkling defect, the present invention judges whether the processed automobile connecting plate has a wrinkling defect, can more accurately detect the wrinkling defect, improve the product quality and production efficiency of the automobile connecting plate, reduce the defective rate and production cost, and is applicable to the on-line detection and quality control of the automobile connecting plate.

Claims

1. An intelligent recognition method for processing automobile connecting plates, characterized in that, Including: The surface image of the corner position of the processed automotive connecting plate; Update the initial low threshold multiple times. Determine the target low threshold according to the change in the number and connection of the black connected regions in the binary images corresponding to the low threshold before and after each update. Perform threshold segmentation on the surface image using the target low threshold, extract the skeleton of the largest black connected region in the obtained low binary image to obtain the guide line contour. Calculate the first probability of the connecting plate having a wrinkling defect according to the curvature change of the contour points on the guide line contour; Update the initial high threshold multiple times. Determine the target high threshold according to the change in the number and connection of the white connected regions in the binary images corresponding to the high threshold before and after each update. Perform threshold segmentation on the surface image using the target high threshold, extract the skeleton of the white connected region in the obtained high binary image to obtain the intersection angle contour of the surface image. Obtain the contour branch points with multiple neighborhood points in the 8-neighborhood of the intersection angle contour, perform chain code extraction on the intersection angle contour based on the contour branch points, and calculate the second probability of the connecting plate having a wrinkling defect according to the difference in the chain code sets of the contour branch points in different directions; Judge whether the processed automotive connecting plate has a wrinkling defect according to the first probability and the second probability.

2. The intelligent recognition method for processing an automotive connecting plate according to claim 1, characterized in that, The step of updating the initial low threshold multiple times and determining the target low threshold according to the change in the number and connection of the black connected regions in the binary images corresponding to the low threshold before and after each update includes: Update the initial low threshold multiple times. Perform threshold segmentation on the surface image according to the low threshold after each update to obtain a binary image corresponding to the low threshold after each update. Perform connected component analysis on all black pixel points in the binary image corresponding to the low threshold after each update to obtain multiple black connected components; until the two black connected components with the largest areas in the binary image corresponding to the low threshold after the th update are merged into one black connected component in the binary image corresponding to the low threshold after the th update, stop updating the initial low threshold, and use the low threshold after the th update as the target low threshold.

3. An intelligent recognition method for processing an automotive connecting plate according to claim 1, characterized in that, The method for obtaining the curvature at the contour point includes: For any contour point on the guide line contour, perform linear fitting on the positions of this contour point and its 6 adjacent points, and use the slope of the obtained straight line as the curvature at this contour point.

4. The intelligent recognition method for processing an automobile connecting plate according to claim 1, wherein, The step of calculating the first probability of the connecting plate having a wrinkling defect according to the curvature change of the contour points on the guide line contour includes: ； In the formula, is the first probability of wrinkling defect in the connecting plate, is the natural exponential function, is the curvature of the th contour point on the guide line profile, is the curvature of the th contour point on the guide line profile, is the number of all contour points included in the guide line profile.

5. An intelligent recognition method for processing automotive connecting plates according to claim 1, characterized in that, The step of updating the initial high threshold multiple times and determining the target high threshold according to the change in the number and connection of the white connected regions in the binary images corresponding to the high threshold before and after each update includes: Update the initial high threshold multiple times. Perform threshold segmentation on the surface image according to the high threshold after each update to obtain a binary image corresponding to the high threshold after each update. Perform connected component analysis on all white pixel points in the binary image corresponding to the high threshold after each update to obtain multiple white connected components; until there is only 1 white connected component in the binary image corresponding to the high threshold after the -th update , and when this white connected component is split into at least 2 white connected components in the binary image corresponding to the high threshold after the -th update, stop updating the initial high threshold and use the high threshold after the -th update as the target high threshold.

6. The intelligent recognition method for processing an automobile connecting plate according to claim 1, characterized in that, The step of performing chain code extraction on the intersection angle contour based on the contour branch points includes: For any contour branch point, along the direction corresponding to any neighborhood point in the 8-neighborhood of this contour branch point, by judging whether there are other contour points in the 8-neighborhood of each contour point, if there are other contour points in the 8-neighborhood, obtain the chain code value of the other contour point compared to this contour point, and sequentially obtain the chain code values of each contour point until there are no other contour points in the 8-neighborhood of the contour point or the other contour points in the 8-neighborhood of the contour point belong to the contour branch point, and form the chain code set of this contour branch point in this direction with all the chain code values.

7. An intelligent recognition method for processing an automotive connecting plate according to claim 1, characterized in that, The step of calculating the second probability of the connecting plate having a wrinkling defect according to the difference in the chain code sets of the contour branch points in different directions includes: ； Wherein, is the second probability of the wrinkling defect existing in the connecting plate, is the number of all contour branch points on the intersection angle contour, is the maximum value of the included angles between the standard chain code values of the th contour branch point in every two different directions, and is the natural exponential function.

8. An intelligent recognition method for processing automotive connection plates according to claim 1, characterized in that, The method for obtaining the standard chain code value of the contour branch point in each direction includes: For any contour branch point, calculate the mean value of all the chain code values in the chain code set of this contour branch point in each direction as the standard chain code value of this contour branch point in each direction.

9. An intelligent recognition method for processing an automobile connecting plate according to claim 1, characterized in that, Determining whether there is a wrinkling defect in the processed automotive connecting plate according to the first probability and the second probability includes: The first probability that the connecting plate has a wrinkling defect and the second probability that the connecting plate has a wrinkling defect where the maximum value is greater than the tolerance it is considered that the connecting plate has a wrinkling defect; Otherwise, calculate the comprehensive probability of the connecting plate having a wrinkling defect based on the first probability and the second probability of the connecting plate having a wrinkling defect. When the comprehensive probability of the connecting plate having a wrinkling defect is greater than the tolerance it is considered that the connecting plate has a wrinkling defect.

10. The intelligent recognition method for processing an automobile connecting plate according to claim 1, characterized in that, The calculation formula for the comprehensive probability of the wrinkling defect existing in the connecting plate is: ； In the formula, is the comprehensive probability of wrinkling defects in the connecting plate, represents the maximum value function, represents the minimum value function; is the first probability of wrinkling defects in the connecting plate, is the second probability of wrinkling defects in the connecting plate.

Images