A leather cutting layout and trajectory generation method based on visual inspection

Abstract

The present invention discloses a leather cutting layout and trajectory generation method based on visual detection, which relates to the field of image segmentation technology and improves the accuracy of leather cutting paths. The present invention obtains the pixel gradient value between each pixel in a two-dimensional image of the leather surface, and then marks a number of leather textures on the front and back sides of the leather structure model, sets a pixel frame to traverse a preset leather custom cutting style image, and obtains the fuzzy weight of each pixel in the pixel frame according to the traversal result, and then performs multi-pixel fuzzification on each pixel according to the fuzzy weight, and classifies each pixel according to the multi-pixel fuzzification result, generates a cutting trajectory path corresponding to the two-dimensional image of the leather surface according to the classification result, and adjusts the vibration range of the vibrating knife by judging whether the cutting trajectory path passes through the leather texture.

Description

Technical Field

[0001] The invention relates to the technical field of image segmentation, and in particular to a leather cutting layout and trajectory generation method based on visual detection. Background Art

[0002] The process of leather cutting and layout mainly involves cutting the leather raw materials into the required shapes and making precise layouts on the leather to maximize the use of leather materials. The traditional method relies on manual measurement, marking and cutting, and workers need to use tools such as rulers and protractors for accurate calibration. This method is not only inefficient, but also prone to human errors, resulting in waste of leather materials and increased production costs.

[0003] With the development of computer vision technology, leather cutting planning through visual inspection has gradually attracted attention. This method uses high-resolution cameras and image processing technology to detect and analyze leather materials in real time to generate optimized cutting layout plans.

[0004] Although visual inspection technology has great advantages in leather cutting and typesetting, it still has the following defects:

[0005] The problem of algorithm universality: Existing layout and trajectory generation algorithms usually generate cutting paths uniformly according to the type of leather, and are unable to generate cutting paths specifically based on the leather's own defects or texture, making it difficult for the cutting process to adapt to complex and changing production environments and needs.

[0006] Disadvantages of material defect detection: Although visual inspection technology can identify some obvious leather defects, it can only detect small, hidden defects, such as defects in the leather itself, which will affect the quality of the final product.

[0007] Therefore, how to accurately detect the leather surface texture or defects and generate a dedicated cutting path for each leather is a difficulty in the prior art. To this end, a leather cutting layout and trajectory generation method based on visual detection is provided. Summary of the invention

[0008] In order to solve the above technical problems, the purpose of the present invention is to provide a leather cutting layout and trajectory generation method based on visual detection.

[0009] In order to achieve the above object, the present invention provides the following technical solutions:

[0010] A method for leather cutting layout and trajectory generation based on visual detection, comprising the following steps:

[0011] Step S1, collecting a plurality of laser image data and optical image data of the leather to be processed from the front and back sides respectively by using a laser camera and an optical camera, and establishing a leather structure model of the leather to be processed according to the laser image data and the optical image data;

[0012] Step S2, extracting a two-dimensional image of the leather surface from the front and back sides of the leather structure model, obtaining the pixel value of each pixel in the two-dimensional image of the leather surface, obtaining the pixel gradient value between each pixel and the pixel at the adjacent position, and then marking a number of leather textures on the front and back sides of the leather structure model according to the pixel gradient value between each pixel value;

[0013] Step S3, setting a pixel frame to traverse a preset leather custom cutting style image, and performing multi-pixel blurring on each pixel according to the distance correlation coefficient and the pixel correlation coefficient between each pixel in the pixel frame and the central pixel according to the traversal result, and classifying each pixel according to the multi-pixel blurring result, and generating a cutting trajectory path corresponding to the leather custom cutting style image according to the classification result;

[0014] Step S4, call the vibrating knife to cut each leather to be processed along the cutting trajectory path, and determine whether the cutting trajectory path passes through the leather texture during the cutting process, and then adjust the vibration range of the vibrating knife according to the thickness relationship between the leather texture and the cutting trajectory path until the cutting trajectory path ends.

[0015] Furthermore, the process of establishing the leather structure model includes:

[0016] Setting a number for each leather to be processed, and then using a laser camera and an optical camera to respectively capture laser image data and optical image data of the front and back sides of the leather to be processed, and marking the laser image data and the optical image data with corresponding numbers;

[0017] A laser image model and an optical image model of the leather to be processed are established according to the laser image data and the optical image data respectively, the laser image model and the optical image model are overlapped along the edge position to obtain a leather structure model, and the leather structure model is marked with a corresponding number.

[0018] Furthermore, the leather texture division process includes:

[0019] Extracting two-dimensional images of the leather surface from the front and back sides of the leather structure model respectively, marking numbers on the two-dimensional images of the leather surface on the front and back sides, establishing a two-dimensional coordinate system, mapping each two-dimensional image of the leather surface onto the two-dimensional coordinate system respectively, obtaining the pixel value of each pixel in each two-dimensional image of the leather surface, and then obtaining the pixel gradient value between each pixel and its adjacent pixels;

[0020] Setting a pixel gradient threshold, and judging whether two pixels are mutually associated pixels or mutually boundary pixels according to the magnitude relationship between the pixel gradient value between two pixels and the pixel gradient threshold;

[0021] Adjacent and mutually associated pixels in each leather surface two-dimensional image are connected in sequence, and then the leather surface two-dimensional image is divided into several associated pixel domains, and the pixels at the junction of the associated pixel domains and adjacent to each other are marked on each leather surface two-dimensional image, and then several leather textures are marked on the leather surface two-dimensional image, and each leather surface two-dimensional image is mapped onto the leather structure model.

[0022] Furthermore, the process of obtaining the distance correlation coefficient and the pixel correlation coefficient includes:

[0023] Grayscale the preset leather custom cutting style image, and obtain the pixel value of each pixel in the leather custom cutting style image after grayscale processing, select a 3*3 pixel frame from the upper left corner and the lower right corner of the leather custom cutting style image, and then establish a two-dimensional coordinate system with the central pixel of the pixel frame as the coordinate origin, and set the coordinate points of other pixels in the pixel frame;

[0024] Then, the distance correlation coefficient α and the pixel correlation coefficient β between each pixel in the pixel frame and the central pixel are obtained;

[0025] Furthermore, the calculation formulas of the distance correlation coefficient α and the distance correlation coefficient β are as follows:

[0026] ;

[0027] ;

[0028] in and They represent the distance correlation coefficient between the jth pixel and the central pixel and the pixel correlation coefficient, σ1 and σ2 are relationship parameters, a, b, q j and p j A and Q are the horizontal and vertical coordinates of the center pixel and the jth pixel, respectively. j Represent the pixel values ​​of the center pixel and the j-th pixel respectively, where j is a natural number greater than 0.

[0029] Furthermore, the process of performing multi-pixel blurring on each pixel includes:

[0030] The fuzzy weight of each pixel is obtained according to the distance correlation coefficient α and the pixel correlation coefficient β, and each pixel is blurred by the fuzzy weight. The calculation formula of each pixel blur is:

[0031] ;

[0032] in represents the pixel value after blurring of the jth pixel, w j represents the fuzzy weight of the jth pixel, where ;

[0033] The pixel frame in the upper left corner of the leather custom cutting style image is traversed to the left in sequence, and the pixel frame in the upper right corner is traversed to the right in sequence. When the pixel frames starting from both sides overlap, the heavy pixel blurring ends, and then pixel frames of 5*5, 7*7, ..., t*t are selected again to perform pixel blurring on the leather custom cutting style image until the pixel frame can no longer select full pixels, and t is an odd number greater than 3.

[0034] Furthermore, the process of obtaining the clipping trajectory path includes:

[0035] Obtain the pixel gradient value between each pixel and its adjacent pixels, and according to the positive and negative signs of the pixel gradient value, obtain the gradient direction of each pixel in different directions, set the primary gradient threshold and the secondary gradient threshold, wherein the primary gradient threshold is greater than the secondary gradient threshold;

[0036] According to the magnitude relationship between the pixel gradient value between each pixel and its adjacent pixels and the primary gradient threshold and the secondary gradient threshold, as well as the number of adjacent pixels of various magnitude relationships, the pixels are marked as outer boundary pixels, isolated pixels, inner boundary pixels or no marking is performed;

[0037] Firstly, the outer boundary pixels and inner boundary pixels with the same gradient direction and being adjacent to each other are sequentially connected to obtain the outer boundary clipping track and the inner boundary clipping track of the preset leather custom cutting style image, and then it is determined whether there are isolated pixels between each outer boundary clipping track and the inner boundary clipping track;

[0038] If so, the isolated pixels are marked as connected pixels, and the outer boundary clipping trajectory and the inner boundary clipping trajectory at the corresponding position are connected through the connected pixels;

[0039] If there are outer boundary pixels or inner boundary pixels that are not connected to any pixels of the same type, and isolated pixels are not changed into connected pixels, the marks of the corresponding pixels are all removed, and then the cutting trajectory path corresponding to the leather custom cutting style image is obtained.

[0040] Further, the process of adjusting the vibration range of the vibrating knife includes:

[0041] The vibrating knife is called to cut each leather to be processed along the cutting trajectory path. Before the vibrating knife starts cutting, it is determined whether the cutting trajectory path passes through the leather texture on the front and back sides of the leather structure model corresponding to the leather to be processed. If not, the vibration range of the vibrating knife is adjusted to the thickness of the cutting trajectory path;

[0042] If it passes, then determine whether the thickness of the corresponding cutting track path portion is greater than or equal to the leather texture. If it is determined that the thickness of the cutting track path portion is greater than or equal to the thickness of the leather texture, then adjust the vibration range of the vibrating knife to the thickness of the cutting track path;

[0043] If the thickness of the cutting trajectory path portion is smaller than the thickness of the leather texture, the vibration range of the vibrating knife is adjusted to the thickness of the leather texture until the cutting trajectory path does not pass through the leather texture or the thickness of the cutting trajectory path portion is greater than or equal to the thickness of the leather texture.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] The present invention obtains the pixel gradient value between each pixel in the two-dimensional image of the leather surface to mark a plurality of leather textures on the front and back sides of the leather structure model, sets a pixel frame to traverse the preset leather custom cutting style image, and performs multi-pixel blurring on each pixel according to the distance correlation coefficient and the pixel correlation coefficient between each pixel in the pixel frame and the central pixel according to the traversal result, and then classifies each pixel according to the multi-pixel blurring result, and generates a cutting trajectory path corresponding to the two-dimensional image of the leather surface according to the classification result, and adjusts the vibration range of the vibrating knife by judging whether the cutting trajectory path passes through the leather texture, thereby realizing the generation of an exclusive cutting path for each leather in a targeted manner on the basis of accurately detecting the leather surface texture. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention.

[0047] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0048] To make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0049] Example 1

[0050] like Figure 1 As shown, a leather cutting layout and trajectory generation method based on visual detection includes the following steps:

[0051] Step S1, collecting a plurality of laser image data and optical image data of the leather to be processed from the front and back sides respectively by using a laser camera and an optical camera, and establishing a leather structure model of the leather to be processed according to the laser image data and the optical image data;

[0052] Step S2, extracting a two-dimensional image of the leather surface from the front and back sides of the leather structure model, obtaining the pixel value of each pixel in the two-dimensional image of the leather surface, obtaining the pixel gradient value between each pixel and the pixel at the adjacent position, and then marking a number of leather textures on the front and back sides of the leather structure model according to the pixel gradient value between each pixel value;

[0053] Step S3, setting a pixel frame to traverse a preset leather custom cutting style image, and performing multi-pixel blurring on each pixel according to the distance correlation coefficient and the pixel correlation coefficient between each pixel in the pixel frame and the central pixel according to the traversal result, and classifying each pixel according to the multi-pixel blurring result, and generating a cutting trajectory path corresponding to the leather custom cutting style image according to the classification result;

[0054] Step S4, call the vibrating knife to cut each leather to be processed along the cutting trajectory path, and determine whether the cutting trajectory path passes through the leather texture during the cutting process, and then adjust the vibration range of the vibrating knife according to the thickness relationship between the leather texture and the cutting trajectory path until the cutting trajectory path ends.

[0055] Example 2

[0056] This embodiment is a further limitation of Embodiment 1, and Step S1 is implemented by the following process:

[0057] Each leather to be processed is numbered a1, a2, a3, ..., a n, and then respectively capture laser image data and optical image data of the front and back sides of the leather to be processed by a laser camera and an optical camera, and mark the laser image data and the optical image data with corresponding numbers, wherein n is a natural number greater than 0;

[0058] A laser image model and an optical image model of the leather to be processed are established according to the laser image data and the optical image data, respectively. Since the laser image model and the optical image model correspond to the same leather to be processed, the objects represented by the two are consistent;

[0059] Then, the laser image model and the optical image model are overlapped along the edge position to obtain a leather structure model, and the leather structure model is marked with a corresponding number.

[0060] Example 3

[0061] This embodiment is a further limitation of Embodiment 1, and Step S2 is implemented by the following process:

[0062] Extract the leather surface two-dimensional images from the front and back sides of the leather structure model respectively, and mark the two-dimensional images of the leather surface on the front and back sides with numbers a i，正 and a i，反 , where i is a natural number greater than 0 and less than or equal to n;

[0063] Establishing a two-dimensional coordinate system, mapping each two-dimensional image of the leather surface onto the two-dimensional coordinate system, obtaining the pixel value of each pixel in each two-dimensional image of the leather surface, and then obtaining the pixel gradient value between each pixel and its adjacent pixels;

[0064] Set a pixel gradient threshold. If the pixel gradient value between two pixels is less than or equal to the pixel gradient threshold, the two corresponding pixels are determined to be related pixels.

[0065] If the pixel gradient value between two pixels is greater than the pixel gradient threshold, it is determined that the corresponding two pixels are boundary pixels;

[0066] The pixels that are adjacent and mutually related in each two-dimensional image of the leather surface are sequentially connected, thereby dividing the two-dimensional image of the leather surface into a plurality of related pixel domains;

[0067] Since the pixel values ​​in the two-dimensional image of the leather surface are either mutually associated pixels or mutually boundary pixels according to the judgment result of the pixel gradient value, and the adjacent mutually associated pixels have a related pixel domain, the pixels at the junction of any two associated pixel domains must be mutually boundary pixels;

[0068] Marking the pixels adjacent to the boundary of the associated pixel domains on each two-dimensional image of the leather surface, and then marking a number of leather textures on the two-dimensional image of the leather surface;

[0069] According to the subscript number of each leather surface two-dimensional image number, each leather surface two-dimensional image is mapped onto the corresponding leather structure model.

[0070] Example 4

[0071] This embodiment is a further limitation of Embodiment 1, and step S3 is implemented by the following process:

[0072] Grayscale the preset leather custom cutting style image, and obtain the pixel value of each pixel in the grayscale leather custom cutting style image, select a 3*3 pixel frame from the upper left corner and the lower right corner of the leather custom cutting style image, and then establish a two-dimensional coordinate system with the center pixel of the pixel frame as the coordinate origin, and set the coordinate points of other pixels in the pixel frame, such as (1, 0), (0, -1), etc.;

[0073] Then, the distance correlation coefficient α and the pixel correlation coefficient β between each pixel in the pixel frame and the central pixel are obtained, wherein the calculation formula of the distance correlation coefficient α is:

[0074] ;

[0075] The calculation formula of the distance correlation coefficient β is: ;

[0076] in and They represent the distance correlation coefficient between the jth pixel and the central pixel and the pixel correlation coefficient, σ1 and σ2 are relationship parameters, a, b, q j and p j A and Q are the horizontal and vertical coordinates of the center pixel and the jth pixel, respectively. j Represent the pixel values ​​of the center pixel and the j-th pixel respectively;

[0077] Then, the fuzzy weight of each pixel is obtained according to the distance correlation coefficient α and the pixel correlation coefficient β, and each pixel is blurred by the fuzzy weight. The calculation formula of each pixel blur is:

[0078] ;

[0079] in represents the pixel value after blurring of the jth pixel, w j represents the fuzzy weight of the jth pixel, where ;

[0080] The pixel frame in the upper left corner of the leather custom cutting style image is traversed to the left in sequence, and the pixel frame in the upper right corner is traversed to the right in sequence. When the pixel frames starting from both sides overlap, the heavy pixel blurring ends, and then pixel frames of 5*5, 7*7, ..., t*t are selected again to perform pixel blurring on the leather custom cutting style image until the pixel frame can no longer select full pixels, and t is an odd number greater than 3.

[0081] Further, the pixel gradient value between each pixel and its adjacent pixels is obtained, and according to the positive and negative signs of the pixel gradient value, the gradient direction of each pixel in different directions is obtained, and a primary gradient threshold and a secondary gradient threshold are set, wherein the primary gradient threshold is greater than the secondary gradient threshold;

[0082] If there is a pixel whose pixel gradient value with more than three adjacent pixels is greater than or equal to the primary gradient threshold, the pixel is marked as an outer boundary pixel;

[0083] If there is a pixel whose pixel gradient value with more than five adjacent pixels is less than the primary gradient threshold, and whose pixel gradient value with more than five adjacent pixels is greater than the secondary gradient threshold, then the pixel is marked as an isolated pixel;

[0084] If the pixel gradient value between a pixel and more than three adjacent pixels is greater than or equal to the secondary gradient threshold, but less than the primary gradient threshold, the pixel is marked as an inner boundary pixel. In other cases, the pixel is not marked.

[0085] Firstly, the outer boundary pixels and inner boundary pixels with the same gradient direction and being adjacent to each other are sequentially connected to obtain the outer boundary clipping track and the inner boundary clipping track of the preset leather custom cutting style image, and then it is determined whether there are isolated pixels between each outer boundary clipping track and the inner boundary clipping track;

[0086] If so, the isolated pixels are marked as connected pixels, and the outer boundary clipping trajectory and the inner boundary clipping trajectory at the corresponding position are connected through the connected pixels;

[0087] If there are outer boundary pixels or inner boundary pixels that are not connected to any pixels of the same type, and if isolated pixels are not changed to connected pixels, all the labels of the corresponding pixels are removed;

[0088] Repeat the above process of obtaining the outer boundary cutting trajectory and the inner boundary cutting trajectory and connecting the two to obtain the cutting trajectory path corresponding to the leather custom cutting style image, and map the cutting trajectory path to the front and back sides of the leather structure model of each leather to be processed.

[0089] Example 5

[0090] This embodiment is a further limitation of Embodiment 1, and step S4 is implemented by the following process:

[0091] The vibrating knife is called to cut each leather to be processed along the cutting trajectory path. Before the vibrating knife starts cutting, it is determined whether the cutting trajectory path passes through the leather texture on the front and back sides of the leather structure model corresponding to the leather to be processed. If not, the vibration range of the vibrating knife is adjusted to the thickness of the cutting trajectory path;

[0092] If it passes, then determine whether the thickness of the corresponding cutting track path portion is greater than or equal to the leather texture. If it is determined that the thickness of the cutting track path portion is greater than or equal to the thickness of the leather texture, then adjust the vibration range of the vibrating knife to the thickness of the cutting track path;

[0093] If the thickness of the cutting path is smaller than the thickness of the leather texture, the vibration range of the vibrating knife is adjusted to the thickness of the leather texture until the cutting path does not pass through the leather texture or the thickness of the cutting path is greater than or equal to the thickness of the leather texture;

[0094] Repeat the above process of adjusting the vibration range of the vibration knife until the cutting trajectory path is executed.

[0095] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for leather cutting layout and trajectory generation based on visual detection, characterized in that: The following steps are involved: Step S1, collecting a plurality of laser image data and optical image data of the leather to be processed from the front and back sides respectively by using a laser camera and an optical camera, and establishing a leather structure model of the leather to be processed according to the laser image data and the optical image data; Step S2, extracting two-dimensional images of the leather surface from the front and back sides of the leather structure model respectively, and obtaining the pixel value of each pixel in the two-dimensional image of the leather surface, obtaining the pixel gradient value between each pixel and the pixel at the adjacent position, and then marking a number of leather textures on the front and back sides of the leather structure model according to the pixel gradient value between each pixel value; Step S3, setting a pixel frame to traverse a preset leather custom cutting style image, and performing multi-pixel blurring on each pixel according to the distance correlation coefficient and the pixel correlation coefficient between each pixel in the pixel frame and the central pixel according to the traversal result, and classifying each pixel according to the multi-pixel blurring result, and then generating a cutting trajectory path corresponding to the leather custom cutting style image according to the classification result; The process of obtaining the clipping trajectory path includes: Obtain the pixel gradient value between each pixel and its adjacent pixels, and according to the positive and negative signs of the pixel gradient value, obtain the gradient direction of each pixel in different directions, set the primary gradient threshold and the secondary gradient threshold, wherein the primary gradient threshold is greater than the secondary gradient threshold; According to the magnitude relationship between the pixel gradient value between each pixel and its adjacent pixels and the primary gradient threshold and the secondary gradient threshold, as well as the number of adjacent pixels of various magnitude relationships, the pixels are marked as outer boundary pixels, isolated pixels, inner boundary pixels or no marking is performed; Connecting adjacent outer boundary pixels and inner boundary pixels with the same gradient direction in sequence to obtain outer boundary clipping tracks and inner boundary clipping tracks of the preset leather custom cutting style image, and then determining whether there are isolated pixels between each outer boundary clipping track and inner boundary clipping track; If so, the isolated pixels are marked as connected pixels, and the outer boundary clipping trajectory and the inner boundary clipping trajectory at the corresponding position are connected through the connected pixels; If there are outer boundary pixels or inner boundary pixels that are not connected to any pixels of the same type, and isolated pixels are not changed into connected pixels, all the marks of the corresponding pixels are removed, and then the cutting trajectory path corresponding to the leather custom cutting style image is obtained; Step S4, calling the vibration knife to cut each leather to be processed along the cutting track path, and judging whether the cutting track path passes through the leather texture during the cutting process, and then adjusting the vibration range of the vibration knife according to the thickness relationship between the leather texture and the cutting track path, until the cutting track path ends; The process of adjusting the vibration range of the vibrating knife includes: The vibrating knife is called to cut each leather to be processed along the cutting trajectory path. Before the vibrating knife starts cutting, it is determined whether the cutting trajectory path passes through the leather texture on the front and back sides of the leather structure model corresponding to the leather to be processed. If not, the vibration range of the vibrating knife is adjusted to the thickness of the cutting trajectory path; If it passes, then determine whether the thickness of the corresponding cutting track path portion is greater than or equal to the leather texture. If it is determined that the thickness of the cutting track path portion is greater than or equal to the thickness of the leather texture, then adjust the vibration range of the vibrating knife to the thickness of the cutting track path; If the thickness of the cutting trajectory path portion is smaller than the thickness of the leather texture, the vibration range of the vibrating knife is adjusted to the thickness of the leather texture until the cutting trajectory path does not pass through the leather texture or the thickness of the cutting trajectory path portion is greater than or equal to the thickness of the leather texture.

2. The method for leather cutting layout and trajectory generation based on visual detection according to claim 1, characterized in that: The process of obtaining the pixel gradient value between the pixels includes: Extracting two-dimensional images of the leather surface from the front and back sides of the leather structure model, respectively, establishing a two-dimensional coordinate system, mapping each two-dimensional image of the leather surface onto the two-dimensional coordinate system, obtaining the pixel value of each pixel in each two-dimensional image of the leather surface, and then obtaining the pixel gradient value between each pixel and its adjacent pixels; A pixel gradient threshold is set, and based on the magnitude relationship between the pixel gradient value between two pixels and the pixel gradient threshold, it is determined whether the pixels are mutually associated pixels or mutually boundary pixels.

3. The method for leather cutting layout and trajectory generation based on visual detection according to claim 2, characterized in that: The leather texture division process includes: Adjacent and mutually associated pixels in each leather surface two-dimensional image are connected in sequence, and then the leather surface two-dimensional image is divided into several associated pixel domains, and the pixels at the junction of the associated pixel domains and adjacent to each other are marked on each leather surface two-dimensional image, and then several leather textures are marked on the leather surface two-dimensional image, and each leather surface two-dimensional image is mapped onto the leather structure model.

4. The method for leather cutting layout and trajectory generation based on visual detection according to claim 3, characterized in that: The process of obtaining the distance correlation coefficient and the pixel correlation coefficient includes: Grayscale the preset leather custom cutting style image, and obtain the pixel value of each pixel in the leather custom cutting style image after grayscale processing, select a 3*3 pixel frame from the upper left corner and the lower right corner of the leather custom cutting style image, and then establish a two-dimensional coordinate system with the central pixel of the pixel frame as the coordinate origin, and set the coordinate points of other pixels in the pixel frame; Then, the distance correlation coefficient α and the pixel correlation coefficient β between each pixel in the pixel frame and the central pixel are obtained.

5. The method for leather cutting layout and trajectory generation based on visual detection according to claim 4, characterized in that: The calculation formulas of the distance correlation coefficient α and the distance correlation coefficient β are as follows: ； ； in and They represent the distance correlation coefficient between the jth pixel and the central pixel and the pixel correlation coefficient, σ1 and σ2 are relationship parameters, a, b, q j and p j A and Q are the horizontal and vertical coordinates of the center pixel and the jth pixel, respectively. j Represent the pixel values ​​of the center pixel and the j-th pixel respectively, where j is a natural number greater than 0.

6. The method for leather cutting layout and trajectory generation based on visual detection according to claim 5, characterized in that: The process of multi-pixel blurring for each pixel includes: The fuzzy weight of each pixel is obtained according to the distance correlation coefficient α and the pixel correlation coefficient β, and each pixel is blurred by the fuzzy weight. The calculation formula of each pixel blur is: ； in represents the pixel value after blurring of the jth pixel, w j represents the fuzzy weight of the jth pixel, where ; The pixel frame in the upper left corner of the leather custom cutting style image is traversed to the left in sequence, and the pixel frame in the upper right corner is traversed to the right in sequence. When the pixel frames starting from both sides overlap, the heavy pixel blurring ends, and then pixel frames of 5*5, 7*7, ..., t*t are selected again to perform pixel blurring on the leather custom cutting style image until the pixel frame can no longer select full pixels, and t is an odd number greater than 3.

Images