A vision-based intelligent detection method and device for punching and cutting defects

Abstract

The present application relates to the technical field of visual detection, and particularly relates to a kind of based on vision's punching piece defect intelligent detection method and device, method steps include: from below to top to piece of cloth lighting, obtain the overhead image of piece of cloth, and after removing texture pattern, black and white processing obtains the image to be measured;Identify the inflection point of piece of cloth outer contour in the image to be measured, obtain the detection contour of piece of cloth;Detection contour is compared with the design contour of piece of cloth, and fold condition is judged;According to fold condition, the stretching image is obtained by stretching detection contour and its internal region, so that detection contour and design contour range coincide;Identify each hole area in the stretching image as detection hole area, and calculate the edge roughness of each detection hole area;According to design contour, determine the design hole area of each hole, calculate the coincidence degree between corresponding design hole area and detection hole area.The present application can effectively improve the detection accuracy and detection efficiency of punching piece defect.

Description

Technical Field

[0001] This invention relates to the field of visual inspection technology, and in particular to a vision-based intelligent detection method and device for defects in punched sheet cuttings. Background Technology

[0002] In the automotive interior manufacturing industry, holes for installation need to be punched into the fabric panels. The quality of these holes directly affects the product's aesthetics and durability. Current inspection methods primarily rely on manual measurement or simple visual inspection methods with threshold divisions. However, these methods have significant shortcomings in practical use: due to the inherent flexibility of the fabric panels, wrinkles inevitably form during movement. These wrinkles can cause the holes to shift or deform, affecting the accuracy of traditional inspection methods. This necessitates carefully flattening every part of the fabric panel during inspection, severely impacting inspection efficiency. Summary of the Invention

[0003] This invention provides a vision-based intelligent detection method and device for defects in punched sheet cuttings, which can effectively solve the problems in the background art.

[0004] This invention provides a vision-based intelligent detection method for defects in punched sheet cuttings, comprising the following steps:

[0005] The fabric piece is illuminated from bottom to top to obtain a grayscale image of its top view. The texture and pattern in the top view image are then removed and the image is blacked out to obtain the image to be tested.

[0006] Identify the inflection points of the outer contour of the cut piece in the image to be tested, and obtain the detected outer contour of the cut piece;

[0007] The outer contour of the detection is compared with the outer contour of the design of the cut piece, and the wrinkle situation is judged. If it exceeds the allowable range, an alarm is triggered; if it is within the allowable range, the stretching area and stretching parameters of the outer contour of the detection are obtained according to the wrinkle situation. Then, the outer contour of the detection and its internal area are stretched to obtain a stretched image, so that the outer contour of the detection and the outer contour of the design overlap.

[0008] The regions of each hole in the stretched image are identified as detection hole regions, and the roughness of the edge of each detection hole region is calculated. Holes with roughness exceeding the set range are judged as burrs.

[0009] The design hole area of ​​each hole is determined based on the outer contour of the design. The overlap between the corresponding design hole area and the inspection hole area is calculated. Holes with an overlap exceeding the set range are judged as having unqualified dimensions.

[0010] Furthermore, the specific steps for removing texture patterns from the top-view image are as follows:

[0011] Perform a Fourier transform on the top-view image to convert it from the spatial domain to the frequency domain;

[0012] Eliminate noise in the frequency domain of top-view images;

[0013] Perform an inverse Fourier transform on the top view image in the frequency domain to convert the top view image back from the frequency domain to the spatial domain;

[0014] The noise in the spatial domain of the top-view image is eliminated to obtain the final image.

[0015] Furthermore, the specific steps for black-and-white conversion of the top-view image to obtain the image to be tested are as follows:

[0016] Divide the image to be tested into N regions of equal size, and calculate the segmentation value Tn = μn - 1.5σn for the nth region;

[0017] Where μn is the average gray value of all pixels in the nth region; σn is the standard deviation of the gray values ​​of all pixels in the nth region;

[0018] Set all pixels in the nth region with a gray value greater than or equal to Tn to black, and set all pixels in the nth region with a gray value less than Tn to white.

[0019] Furthermore, the specific steps for identifying the inflection points of the outer contour of the cut piece in the image to be tested, and obtaining the detected outer contour of the cut piece, are as follows:

[0020] Calculate the distance between each inflection point and other inflection points on the outer contour of the design to form multiple comparison groups;

[0021] Starting from the leftmost column of the image to be tested, search column by column until the first black pixel is found, and mark it as the origin O.

[0022] Traverse all black pixels in the remaining images to be tested, and count the colors of the 8 pixels surrounding each pixel. If the number of white pixels is greater than the number of black pixels, then record that pixel as a point to be determined.

[0023] Calculate the distance between each point to be determined and the origin, and compare it with multiple comparison groups to find the closest comparison group, thereby determining the correspondence between the origin and the points to be determined and the various inflection points of the designed outer contour;

[0024] The origin and the undetermined point in the image to be tested are connected in the order of the inflection points in the designed outer contour to obtain the detected outer contour of the cut piece.

[0025] Furthermore, the specific steps for obtaining the pleat condition of the cut piece are as follows:

[0026] Calculate the minimum bounding box center point of the detected outer contour and the designed outer contour, and make the two centers coincide by translation, and make the two placement directions the same by rotation;

[0027] Calculate the degree of overlap between the detected outer contour and the designed outer contour.

[0028] Furthermore, based on the wrinkle pattern, the stretching region and stretching parameters of the image to be tested are obtained, and then the image to be tested is stretched to obtain the stretched image. The specific steps are as follows:

[0029] Identify and detect non-straight line segments on the outer contour, and fit an arc line segment for each non-straight line segment;

[0030] For each arc segment, the straight line passing through the center of the arc segment and the center point is taken as the center line. The angle between the center line and the corresponding outer contour line segment of the arc segment is recorded as the offset angle. The fold length is calculated based on the offset angle and the chord height of the arc segment. Then, the point obtained by moving the fold length inward from the center point of the arc segment along the center line is recorded as the fold vertex. Connecting the fold vertex and the two ends of the arc segment forms a triangular region. The part of the outer contour that is located within the triangular region is a stretching region.

[0031] All pixels within the stretching area are moved according to their distance from the corresponding center line and arc segment to achieve the stretching effect, and interpolation is used to fill in the missing pixels within the stretching area.

[0032] Furthermore, the specific steps for determining whether a hole has burrs are as follows:

[0033] For each detection hole area, calculate the minimum bounding box center point of the detection hole area, calculate the distance between all edge points of the hole area and the center point, and calculate the standard deviation of all distance values. The standard deviation is recorded as the roughness.

[0034] Furthermore, the step of determining whether a hole has burrs also includes:

[0035] The edge point sequence is converted to a polar coordinate system, with the center point of the minimum bounding box located at the origin of the polar coordinate system, generating a radius-angle curve;

[0036] The number of local extrema of the detection curve is recorded, and the angle of each local extrema is recorded.

[0037] The number of local extreme points and their corresponding angles recorded from multiple tests are compared to determine the burr inheritance status.

[0038] Furthermore, prior to testing, a brightness calibration step is included, which includes:

[0039] Capture blank background images without cropping to detect the uniformity of illumination distribution;

[0040] If the difference between the brightness of the central area and the edge exceeds the set threshold, the light source angle will be adjusted or the brightness will be compensated until the difference returns to within the set threshold.

[0041] After calibration, the light source parameters are locked to ensure consistency in subsequent image acquisition.

[0042] This invention also provides a vision-based intelligent detection device for defects in punched sheet cuttings, comprising:

[0043] The workbench is made of transparent material and is used to place the cut pieces.

[0044] The light source is located below the workbench;

[0045] The camera is positioned above the workbench;

[0046] The processor is used to implement the aforementioned vision-based intelligent detection method for defects in punched pieces.

[0047] The technical solution of this invention can achieve the following technical effects:

[0048] This method precisely divides the image of the cut piece by setting a stretching area, and then stretches and corrects the captured image of the cut piece, thereby eliminating the hole deformation and misalignment problems caused by the folds of the cut piece in the image, thus making the detection of holes more accurate; and it also eliminates the need for personnel to carefully lay the cut piece flat during the detection, thus effectively improving the detection efficiency. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart of the vision-based intelligent detection method for defects in punched pieces in this invention;

[0051] Figure 2 This is a schematic diagram of the outer contour of the cut piece in this invention;

[0052] Figure 3 This is a schematic diagram of the outer contour of the cut piece in this invention;

[0053] Figure 4 This is an enlarged view showing the various possibilities of the inflection point on the graph in this invention;

[0054] Figure 5This is an enlarged view showing the various possibilities of the edge points being located on the graph in this invention. Detailed Implementation

[0055] The basic principles and main features of the technical solution of the present invention will be described below with reference to the accompanying drawings of the embodiments of the present invention. The following description will use one or more embodiments for a more intuitive understanding. These embodiments are merely some, not all, of the embodiments of the present invention.

[0056] In the description of this invention, the terms indicating orientation or positional relationship (such as up, down, left, right, etc.) are based on the orientation shown in the drawings or some conventional positional relationships, and are only for the convenience of describing this invention and simplifying the description, and are not intended to indicate or imply that the features referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting this invention.

[0057] This invention relates to a vision-based intelligent detection method for defects in punched sheet cuttings. The main steps include image acquisition, image processing, and calculation. The specific processes of each step are as follows:

[0058] Image acquisition steps: Illuminate the fabric piece from bottom to top. From a top-down perspective, the fabric piece will appear darker, while the holes and areas outside the fabric piece will appear brighter. This will create a top-down image with very clear contrast. Then, convert the top-down image of the fabric piece into a grayscale image. Due to the woven texture commonly found on the surface of decorative fabric materials, these textures will form regularly distributed texture patterns in the image, such as evenly spaced white spots or striped spots. These texture patterns will affect subsequent calculations, so it is necessary to remove the texture patterns from the top-down image. After that, the grayscale image is converted to black and white to further highlight the fabric piece in the image. This black and white image is used as the final image to be tested.

[0059] Image processing steps: This step mainly involves identifying the cropped portions in the image to be tested and eliminating wrinkles in the image through local stretching. Specifically:

[0060] Identify the inflection points of the outer contour of the cut piece in the image to be tested, and then connect the inflection points in sequence to obtain the detected outer contour of the cut piece;

[0061] The outer contour of the detected piece is compared with the outer contour of the design of the piece. The design outer contour is the size of the piece in the design. When the piece has wrinkles, the area of ​​the piece will be smaller than the size of the design when viewed from above. Therefore, this comparison can be used to reflect the wrinkles of the piece. Then, an allowable range can be set, and the wrinkle situation can be judged. If it exceeds the allowable range, it means that the wrinkles of the cut piece are too large, which will cause serious deformation of the hole shape from the top view. The hole shape cannot be restored by stretching the image alone, so an alarm is triggered, and the cut fabric needs to be manually repositioned. If it is within the allowable range, it means that the wrinkles of the cut piece are not large, and the hole shape can be restored by stretching the image. Then, the stretching area and stretching parameters of the detection outer contour are obtained according to the wrinkle situation. Then, the detection outer contour and its internal area are stretched to obtain a stretched image. The stretching area is the area that will deform during stretching. Other areas will be moved by translation and rotation to ensure that they are not deformed. The stretching parameters are the data used during stretching, such as the direction and distance of movement of each pixel in the stretching area. After stretching, the range of the detection outer contour and the design outer contour will overlap. At this time, the hole located in the wrinkle area (i.e., the stretching area) will also be stretched and restored to a hole shape that is closer to the real shape, and its position on the image is also corrected. Only then can the detection outer contour and the design outer contour be compared.

[0062] Finally, the calculation step involves determining whether there are defects in the punched holes on the cut piece based on the stretched image after stretching and correction. The details are as follows:

[0063] First, it's necessary to determine if there are burrs on the edges of the holes. In the stretched image, white represents the part that allows light to pass through, which is the hole. We only need to find the area formed by adjacent white pixels to identify a hole region. Using this method, we can identify the areas of each hole in the stretched image as the detection hole regions. After the hole regions are identified, the roughness of the edges of each detection hole region is calculated. Roughness mainly indicates whether the detection hole region is smooth; the smoother the hole edge, the lower the roughness. Holes with roughness exceeding a set range are judged as burrs.

[0064] Next, after determining whether the size and position of the holes meet the design requirements, the design hole area for each hole is determined based on the outer contour of the design. The design hole area is a virtual area. The closer the size and position of the hole are to the design size and position, the more the detection hole area should overlap with the design hole area. That is, the number of white pixels representing the hole in the design hole area should be more, while the number of white pixels around the outside of the design hole area should be less. Based on this idea, the overlap between the corresponding design hole area and the detection hole area can be calculated, and holes with an overlap exceeding the set range are judged as having unqualified dimensions.

[0065] As can be seen, this method precisely divides the image of the cut piece by setting a stretching area, and then stretches and corrects the captured image of the cut piece, thereby eliminating the hole deformation and misalignment problems caused by the folds of the cut piece in the image, thus making the detection of holes more accurate; and the cut piece does not need to be carefully laid flat by personnel during the detection, thus effectively improving the detection efficiency.

[0066] Preferably, the specific steps for removing texture patterns from a top-view image are as follows:

[0067] Performing a Fourier transform on the top-view image converts it from the spatial domain to the frequency domain. In the frequency domain, image information can be processed from the perspective of frequency. In the frequency domain, features with obvious periodicity, such as woven texture patterns, will be transformed into noise far from the edges. At this time, it is only necessary to use filtering algorithms and other methods to eliminate the noise in the frequency domain of the top-view image, and the texture pattern can be removed in the frequency domain.

[0068] Next, an inverse Fourier transform is performed on the top view image in the frequency domain to convert the top view image back to the spatial domain. This way, the image in the spatial domain will no longer display texture patterns. However, this will generate some additional noise, and the original image may also have some noise due to impurities. Therefore, a filtering algorithm is needed to remove the noise in the spatial domain to obtain the final image. The final image will not have any noise that interferes with subsequent calculations.

[0069] Preferably, the specific steps for obtaining the image to be tested by black and white conversion of the top-view image are as follows:

[0070] Divide the image to be tested into N regions of equal size, and calculate the segmentation value Tn = μn - 1.5σn for the nth region;

[0071] Where μn is the average gray value of all pixels in the nth region, used to reflect local brightness; σn is the standard deviation of the gray values ​​of all pixels in the nth region, used to measure contrast; 1.5 is an adjustment coefficient, which can be a fixed value or adjusted according to the specific detection scenario requirements;

[0072] Set all pixels in the nth region with a gray value greater than or equal to Tn to black, and set all pixels in the nth region with a gray value less than Tn to white.

[0073] Preferably, the specific steps for identifying the inflection points of the outer contour of the cut piece in the image to be tested and obtaining the detected outer contour of the cut piece are as follows:

[0074] Calculate the distance between each inflection point and other inflection points on the outer contour of the design, forming multiple comparison groups, such as... Figure 2As shown, assuming the outer contour of the cut piece is a quadrilateral ABCD, then points A, B, C, and D are used as the base points to calculate the distances between each point and the other points, ultimately forming four comparison groups: [AB, AC, AD], [BA, BC, BD], [CA, CB, CD], and [DA, DB, DC]. These values ​​are calculated based on the dimensions of the cut piece design, and therefore remain fixed during the inspection of the same type of cut piece.

[0075] The next step is to process the captured image of the cut piece to be tested, i.e., the image to be tested, to identify the position of the cut piece in the image to be tested and form the detection outer contour:

[0076] First, start searching column by column from the leftmost edge of the image to be tested. Since noise has already been removed from the image in the previous steps, the first black pixel found will be a corner inflection point on the cropped image. Figure 3 As shown, this point is denoted as the origin O;

[0077] Iterate through all black pixels in the remaining test images, and count the colors of the 8 pixels surrounding each black pixel. There will be significant differences in the proportion of black and white pixels at midpoints, inflection points, and edge points. At midpoints, all 8 pixels surrounding the black pixel will be black; at inflection points, the situation will be as follows... Figure 4 As shown, the number of white pixels is greater than the number of black pixels; the situation for edge points is as follows. Figure 5 As shown, the number of white pixels is less than or equal to the number of black pixels. Therefore, if a black pixel is detected where, among its eight surrounding pixels, the number of white pixels is greater than the number of black pixels, then this pixel may be an inflection point. Of course, misjudgments may occur when there is a prominent jagged edge or wind field, for example... Figure 3 As shown in the figure, point H is the only point that can be recorded as a point to be determined, pending further judgment.

[0078] Then calculate the distance between each point to be determined and the origin, for example, using... Figures 2-3 For example, now that the undetermined points E, F, G, and H have been found, calculate OE, OF, OH, and OG, and compare them with multiple comparison groups ([AB, AC, AD], [BA, BC, BD], [CA, CB, CD], and [DA, DB, DC]) to find the closest comparison group. This determines the correspondence between the origin and the undetermined points and the various inflection points of the designed outer contour. For example, if we can find that the combination [OE, OF, OG] corresponds to [AB, AC, AD], then it means that the correspondence between the designed outer contour and the detected outer contour is A corresponds to O, B corresponds to E, C corresponds to F, and D corresponds to G. The undetermined points E, F, and G are inflection points, while point H is a misjudged point.

[0079] Then, the origin O and the points E, F, and G in the image to be tested can be connected according to the connection order ABCD of the inflection points in the designed outer contour, thus obtaining the detected outer contour OEFG of the cut piece. This not only allows for the rapid differentiation between the true inflection points and the misjudged points in the image to be tested, but also accurately determines the correspondence between the designed outer contour and the detected outer contour. In addition to shape and size, even position can be quickly determined, facilitating subsequent comparison between the two.

[0080] Preferably, the specific steps for obtaining the wrinkle condition of the cut piece are as follows:

[0081] Calculate the minimum bounding box center point of the detected outer contour and the designed outer contour, and make their centers coincide by translation, and make their placement orientation the same by rotation, for example. Figures 2-3 The angles of the cut pieces ABCD and OEFG on the diagram are different, so it is necessary to calculate their angle difference and then rotate one of them so that they are placed in the same direction on the diagram.

[0082] Calculate the overlap between the detected outer contour and the designed outer contour. If there are fewer white pixels inside the edge of the designed outer contour and fewer black pixels outside the edge, then the overlap between the detected outer contour and the designed outer contour is higher, which means that the number of folds in the cut piece detected is less.

[0083] Preferably, the specific steps for obtaining the stretchable region and stretching parameters of the image to be tested based on the wrinkle condition, and then stretching the image to be tested to obtain the stretched image are as follows:

[0084] Identify and detect non-linear line segments on the outer contour, and use the least squares circle fitting method to fit the arc line segment of each non-linear line segment;

[0085] For each arc segment, the straight line passing through the center of the arc segment and its center point is taken as the center line. The center line can be regarded as the central symmetry line of the fold. The angle between the center line and the corresponding outer contour line segment of the arc segment is recorded as the offset angle. The fold length is calculated based on the offset angle and the chord height of the arc segment. Then, the point obtained by moving the fold length inward from the center point of the arc segment along the center line is recorded as the fold vertex. Connecting the fold vertex and the two ends of the arc segment forms a triangular region. The part of the outer contour located within the triangular region is a stretching region. Correspondingly, stretching parameters such as offset angle and chord height are also obtained to calculate the displacement of each pixel.

[0086] The specific formula for calculating the displacement of each pixel within the stretching region is as follows:

[0087] ;

[0088] Where △(x,y) represents the displacement of the pixel in the x-th row and y-th column; h is the chord height of the arc segment; θ is the offset angle; Dcen is the distance between the pixel and the central symmetry line; Dedg is the distance between the pixel and the corresponding edge of the detected outer contour; α is the adjustment coefficient, usually taken as 0.1.

[0089] All pixels within the stretching area are moved according to their distance from the corresponding center line and arc segment to achieve the stretching effect. Since the stretching area becomes larger after stretching, the original number of pixels is insufficient to fill the entire stretched area. Therefore, interpolation is needed to fill in the missing pixels within the stretching area.

[0090] The specific steps for determining whether a hole has burrs are as follows:

[0091] For each detection hole region, calculate the minimum bounding box center point of that region, and calculate the distances between all edge points and the center point. Also calculate the standard deviation of all distance values. The smaller the standard deviation, the more consistent the distances between the edge points and the center point, and the closer the edge of the detection hole region is to a circle. Conversely, the larger the standard deviation, the less circular the edge of the detection hole region is. In this way, the standard deviation can reflect the condition of the detection hole region's edge, and can be recorded as roughness. Of course, other influencing parameters can be added if necessary, and then calculated together with the standard deviation to form the final roughness.

[0092] Preferably, the step of determining whether a hole has burrs also includes:

[0093] The edge point sequence is converted to a polar coordinate system, with the center point of the minimum bounding box located at the origin of the polar coordinate system, generating a radius-angle curve;

[0094] The number of local extreme points of the detection curve is recorded, and the angle of the local extreme points is recorded. This means that the number and location of burrs at the edge of the detection hole area are recorded.

[0095] By comparing the number of local extreme points and their corresponding angles recorded from multiple tests, the burr inheritance status can be determined. For example, if the number and position of burrs on multiple cut pieces do not change, it may indicate that the punching tool is worn. If the number of burrs on multiple cut pieces does not change, but the position changes, it may indicate that the punching tool is loose. This burr inheritance status can be used to determine whether there is a problem in the processing equipment, making it easier for personnel to make quick adjustments.

[0096] Preferably, before testing, a brightness calibration step is also included, the step comprising:

[0097] Capture blank background images without cropping to detect the uniformity of light distribution and prevent some areas from being too dark;

[0098] If the difference between the brightness of the central area and the edge exceeds the set threshold, the light source angle will be adjusted or the brightness will be compensated until the difference returns to within the set threshold.

[0099] After calibration, the light source parameters are locked to ensure consistency in subsequent image acquisition.

[0100] This invention also relates to a vision-based intelligent detection device for defects in punched sheet cuttings, comprising:

[0101] The workbench is made of transparent material and is used to place the cut pieces.

[0102] The light source is located below the workbench; preferably, the light source is an LED light that emits more uniform light.

[0103] The camera is positioned above the worktable;

[0104] The processor is used to implement the aforementioned vision-based intelligent detection method for defects in punched pieces.

[0105] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A vision-based intelligent detection method for defects in punched sheet cutting, characterized in that the steps include... include: The fabric piece is illuminated from bottom to top to obtain a grayscale image of its top view. The texture and pattern in the top view image are then removed and the image is blacked out to obtain the image to be tested. Identify the inflection points of the outer contour of the cut piece in the image to be tested, and obtain the detected outer contour of the cut piece; The outer contour of the detection is compared with the outer contour of the design of the cut piece, and the wrinkle situation is judged. If it exceeds the allowable range, an alarm is triggered; if it is within the allowable range, the stretching area and stretching parameters of the outer contour of the detection are obtained according to the wrinkle situation. Then, the outer contour of the detection and its internal area are stretched to obtain a stretched image, so that the outer contour of the detection and the outer contour of the design overlap. The regions of each hole in the stretched image are identified as detection hole regions, and the roughness of the edge of each detection hole region is calculated. Holes with roughness exceeding the set range are judged as burrs. Determine the design hole area for each hole based on the outer contour of the design, calculate the overlap between the corresponding design hole area and the inspection hole area, and determine the hole whose overlap exceeds the set range as dimensionally unqualified. The specific steps for identifying the inflection points of the outer contour of the cut piece in the image to be tested, and obtaining the detected outer contour of the cut piece, are as follows: Calculate the distance between each inflection point and other inflection points on the outer contour of the design to form multiple comparison groups; Starting from the leftmost column of the image to be tested, search column by column until the first black pixel is found, and mark it as the origin O. Traverse all black pixels in the remaining images to be tested, and count the colors of the 8 pixels surrounding each pixel. If the number of white pixels is greater than the number of black pixels, then record that pixel as a point to be determined. Calculate the distance between each point to be determined and the origin, and compare it with multiple comparison groups to find the closest comparison group, thereby determining the correspondence between the origin and the points to be determined and the various inflection points of the designed outer contour; The origin and the point to be determined in the image to be tested are connected in the order of the inflection points in the designed outer contour to obtain the detected outer contour of the piece. The specific steps to obtain the pleat condition of the cut piece are as follows: Calculate the minimum bounding box center point of the detected outer contour and the designed outer contour, and make the two centers coincide by translation, and make the two placement directions the same by rotation; Calculate the degree of overlap between the detected outer contour and the designed outer contour; The specific steps for obtaining the stretched region and stretching parameters of the image to be tested based on the wrinkle pattern, and then stretching the image to obtain the stretched image are as follows: Identify and detect non-straight line segments on the outer contour, and fit an arc line segment for each non-straight line segment; For each arc segment, the straight line passing through the center of the arc segment and the center point is taken as the center line. The angle between the center line and the corresponding outer contour line segment of the arc segment is recorded as the offset angle. The fold length is calculated based on the offset angle and the chord height of the arc segment. Then, the point obtained by moving the fold length inward from the center point of the arc segment along the center line is recorded as the fold vertex. Connecting the fold vertex and the two ends of the arc segment forms a triangular region. The part of the outer contour that is located within the triangular region is a stretching region. All pixels within the stretching area are moved according to their distance from the corresponding center line and arc segment to achieve the stretching effect, and interpolation is used to fill in the missing pixels within the stretching area.

2. The vision-based intelligent detection method for defects in punched pieces according to claim 1, characterized in that, The specific steps for removing texture patterns from a top-view image are as follows: Perform a Fourier transform on the top-view image to convert it from the spatial domain to the frequency domain; Eliminate noise in the frequency domain of top-view images; Perform an inverse Fourier transform on the top view image in the frequency domain to convert the top view image back from the frequency domain to the spatial domain; The noise in the spatial domain of the top-view image is eliminated to obtain the final image.

3. The vision-based intelligent detection method for defects in punched pieces according to claim 1, characterized in that, The specific steps for obtaining the image to be tested by black and white conversion of the top view image are as follows: Divide the image to be tested into N regions of equal size, and calculate the segmentation value Tn = μn - 1.5σn for the nth region; Where μn is the average gray value of all pixels in the nth region; σn is the standard deviation of the gray values ​​of all pixels in the nth region; Set all pixels in the nth region with a gray value greater than or equal to Tn to black, and set all pixels in the nth region with a gray value less than Tn to white.

4. The vision-based intelligent detection method for defects in punched pieces according to claim 1, characterized in that, The specific steps for determining whether a hole has burrs are as follows: For each detection hole area, calculate the minimum bounding box center point of the detection hole area, calculate the distance between all edge points of the hole area and the center point, and calculate the standard deviation of all distance values. The standard deviation is recorded as the roughness.

5. The vision-based intelligent detection method for defects in punched pieces according to claim 4, characterized in that, The steps for determining whether a hole has burrs also include: The edge point sequence is converted to a polar coordinate system, with the center point of the minimum bounding box located at the origin of the polar coordinate system, generating a radius-angle curve; The number of local extrema of the detection curve is recorded, and the angle of each local extrema is recorded. The number of local extreme points and their corresponding angles recorded from multiple tests are compared to determine the burr inheritance status.

6. The vision-based intelligent detection method for defects in punched pieces according to claim 1, characterized in that, Before testing, a brightness calibration step is also included, which includes: Capture blank background images without cropping to detect the uniformity of illumination distribution; If the difference between the brightness of the central area and the edge exceeds the set threshold, the light source angle will be adjusted or the brightness will be compensated until the difference returns to within the set threshold. After calibration, the light source parameters are locked to ensure consistency in subsequent image acquisition.

7. A vision-based intelligent detection device for defects in punched sheet fabric, characterized in that, include: The workbench is made of transparent material and is used to place the cut pieces. The light source is located below the worktable; The camera is positioned above the worktable; A processor for implementing the vision-based intelligent detection method for defects in punched pieces as described in any one of claims 1 to 6.

Images