A method and device for counting and quality detection of warp-knitted jacquard uppers

Abstract

The application provides a warp-knitted shoe upper quantity counting and quality detection method and device, which is used in the field of warp-knitted shoe upper monitoring and includes the following steps: collecting monitoring video on a shoe production line at a warp-knitted shoe upper completion station; identifying and digitally marking warp-knitted shoe upper in the monitoring video based on target detection technology and video processing technology; establishing a shoe upper finite element model based on standard warp-knitted shoe upper; correcting the three-dimensional parameter range of the finite element model based on material and industrial error; performing three-dimensional scanning on the warp-knitted shoe upper to be tested by using the scanning principle to obtain three-dimensional parameters; matching and determining the three-dimensional parameters with the three-dimensional parameter range of the corrected finite element model; and updating the content of the digital mark based on the matching and determination result. The application realizes warp-knitted shoe upper quantity counting and quality detection based on machine vision technology, establishes a shoe upper finite element model, and improves the quality detection precision and intelligent level.

Description

Technical Field

[0001] This invention belongs to the field of shoe upper monitoring technology, and in particular relates to a method and equipment for quantity statistics and quality testing of warp-knitted jacquard shoe uppers. Background Technology

[0002] Shoes are necessities in people's daily lives. With the continuous development of the manufacturing industry, shoe-making equipment has also made leaps and bounds. However, the existing shoe-making process is still very complex, and the degree of automation is very low. A single shoe requires dozens of processes and hundreds of workers. These manual production procedures directly affect production efficiency, most notably in the quantity counting and quality inspection procedures for warp-knitted jacquard shoe uppers. For example, in the ordinary shoe upper inspection process, workers need to count the quantity and measure the size of the shoe uppers. Currently, this process is entirely done manually. This traditional manual inspection method results in slow speed, human visual errors, and high costs. For warp-knitted jacquard shoe uppers, because they are knitted products, the quantity counting and quality inspection are even more complex than for ordinary leather uppers.

[0003] Therefore, those skilled in the art urgently need to provide methods for quantity statistics and quality inspection, primarily for warp-knitted jacquard shoe uppers, in order to improve the automation level and work efficiency of the existing shoe upper manufacturing industry. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a method and equipment for counting the quantity and quality of warp-knitted jacquard shoe uppers, so as to solve the technical problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for quantity counting and quality inspection of warp-knitted jacquard shoe uppers, including:

[0007] Video monitoring was collected at the finishing station of warp-knitted jacquard shoe uppers on the shoe manufacturing production line.

[0008] Based on target detection technology and video processing technology, the warp-knitted jacquard shoe uppers in the monitoring video are identified and digitally marked;

[0009] A finite element model of the shoe upper was established based on the standard warp-knitted jacquard shoe upper;

[0010] Three-dimensional parameter range of finite element model based on material and industrial error correction;

[0011] Using scanning principles, a three-dimensional scan of the warp-knitted jacquard shoe upper to be tested was performed to obtain three-dimensional parameters;

[0012] The three-dimensional parameters are matched and determined with the range of the three-dimensional parameters of the corrected finite element model;

[0013] The content of the digital marker is updated based on the matching result.

[0014] Preferably, the warp-knitted jacquard shoe uppers in the monitoring video are identified and digitally marked based on target detection technology and video processing technology, specifically including:

[0015] Obtain a training sample set for the detection of warp-knitted jacquard shoe uppers; establish a target detection network, and train the target detection network using the training sample set until the target detection results reach a predetermined standard;

[0016] The monitoring video is processed into frames; the frame images of a continuous fixed time period are used as network inputs and input into the target detection network, and the recognition result of the warp-knitted jacquard shoe upper is output;

[0017] Calculate the influence weights of shoe surface curvature and light intensity on the monitored video pixels; and use the influence weights to optimize the recognition results.

[0018] The optimized identification results are sorted according to monitoring time, and the sorting order is marked with numbers to obtain the content of the number markings;

[0019] The digital markers include: monitoring time, which warp-knitted jacquard upper, its category information and its sorting position within that category, and whether there are any defects.

[0020] Preferably, it also includes determining the statistical results based on the photoelectric effect;

[0021] The photoelectric effect is used to detect whether the warp-knitted jacquard shoe upper passes through the production line, and trigger counting is performed based on the detection results;

[0022] The trigger count result is compared with the digital tag result to obtain the determination digital tag result.

[0023] Preferably, the range of three-dimensional parameters of the finite element model is corrected based on material and industrial errors, specifically including:

[0024] To obtain the maximum elasticity of different knitted materials;

[0025] The material influence weights of the three-dimensional parameters of the finite model are calculated based on the maximum elasticity of the different knitted materials.

[0026] Using real-time knitted materials as the object, the deformation data of the warp-knitted jacquard shoe upper in different directions during circumferential stretching were extracted by measuring the maximum deformation of the warp-knitted jacquard shoe upper.

[0027] The range of the three-dimensional parameters of the finite element model is determined by using the material influence weights and the deformation data, thereby completing the correction of the range of the three-dimensional parameters of the finite element model.

[0028] Preferably, the target detection network structure is as follows:

[0029] Input layer - first convolutional layer - feature extraction layer - fully connected layer - second convolutional layer - feature extraction layer - fusion layer - output layer.

[0030] Preferably, the process of obtaining information on whether there are defects in the content of the digital marker specifically includes:

[0031] In the target detection network, the first convolutional layer performs convolution processing on the image of a flawless warp-knitted shoe upper, constructs a fitness function using the Fisher criterion, and calculates the function value corresponding to each particle;

[0032] The filtering parameters in the target detection network are optimized using the quantum behavior particle swarm optimization algorithm to obtain the optimal filtering parameters.

[0033] Using the optimal filtering parameters, the second convolutional layer performs convolution processing on the framed images, and then performs binarization processing to obtain the defect information in the framed images.

[0034] Preferably, the fitness function expression is:

[0035] Where f(x) represents the function value corresponding to each particle; β is the constant coefficient of the function representing different image sizes; h(x) represents the mean energy of the flawless warp-knitted upper image after convolution processing; and g(y) represents the standard deviation of the flawless warp-knitted upper image after convolution processing.

[0036] A terminal device includes a memory, a processor, and a warp-knitted jacquard shoe upper quantity counting and quality inspection program stored in the memory and executable on the processor. When the processor executes the warp-knitted jacquard shoe upper quantity counting and quality inspection program, it implements the steps of a warp-knitted jacquard shoe upper quantity counting and quality inspection method.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] This invention utilizes machine vision technology to achieve quantity statistics and quality inspection of warp-knitted jacquard shoe uppers. Furthermore, during quality inspection, to enhance intelligence and precision, a finite element model based on the shoe upper is established for further quality assessment. The entire process requires no human intervention, and all warp-knitted shoe uppers on the production line are statistically analyzed and displayed in real time, providing all necessary information and improving the intelligence level of warp-knitted shoe upper processing, thus increasing work efficiency. Attached Figure Description

[0039] 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 embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0040] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] Warp-knitted jacquard upper technology is a solution specifically designed to address the inherent trade-off between the strength and lightweight properties of shoe uppers. It leverages big data technology to apply the most suitable characteristics to each knitting thread. In practical applications, manufacturers have high requirements for the physical properties of the fabric, demanding high tensile strength, tear strength, abrasion resistance, and folding resistance. However, due to the nature of chemical fiber raw materials, the tensile strength, tear strength, Martindale abrasion resistance, and folding resistance of mesh fabrics become unstable after dyeing and finishing processes, making it impossible to guarantee that they meet the brand's physical property requirements, creating high hidden risks for enterprises. This also causes significant confusion for dyeing and finishing plants. Warp-knitted fabrics of the same variety, on the same machine, and with the same process data can exhibit different physical property indicators after high-temperature, high-pressure dyeing, pre-setting, and re-setting processes. Even yarns of the same specifications from the same manufacturer may not meet requirements if they are from different batches. Therefore, current technology involves using raw yarns for trial production, continuously developing small-scale production by selecting similar yarns of different specifications or from different manufacturers, using the same process or equipment, and repeatedly testing after dyeing and finishing to select the optimal solution.

[0043] Example 1:

[0044] This embodiment 1 discloses a method for quantity counting and quality inspection of warp-knitted jacquard shoe uppers, including:

[0045] Video monitoring was collected at the finishing station of warp-knitted jacquard shoe uppers on the shoe manufacturing production line.

[0046] Based on target detection technology and video processing technology, warp-knitted jacquard shoe uppers in monitoring videos are identified and digitally marked;

[0047] A finite element model of the shoe upper was established based on the standard warp-knitted jacquard shoe upper;

[0048] Three-dimensional parameter range of finite element model based on material and industrial error correction;

[0049] Using scanning principles, a three-dimensional scan of the warp-knitted jacquard shoe upper to be tested was performed to obtain three-dimensional parameters;

[0050] The three-dimensional parameters are matched and determined with the range of three-dimensional parameters of the corrected finite element model.

[0051] Update the content of the numeric markers based on the matching results.

[0052] Specifically:

[0053] Based on target detection technology and video processing technology, the warp-knitted jacquard shoe uppers in the monitoring video are identified and digitally labeled, specifically including:

[0054] Obtain a training sample set for warp-knitted jacquard shoe upper detection; establish a target detection network, and train the target detection network using the training sample set until the target detection results reach the predetermined standard; the training sample set includes images of warp-knitted shoe uppers with different knitting patterns, and their corresponding names and categories;

[0055] The monitoring video is processed into frames to obtain frame images; the frame images of a continuous fixed time period are used as input to the target detection network, and the output is the recognition result of the warp-knitted jacquard shoe upper.

[0056] Calculate the influence weights of shoe surface curvature and light intensity on the pixels of the monitoring video; and use the influence weights to optimize the recognition results.

[0057] The optimized recognition results are sorted according to monitoring time, and the sorting order is marked with numbers to obtain the content of the numerical markings;

[0058] The numerical markers include: monitoring time, the number of the warp-knitted jacquard upper, its category information and ranking within that category, and whether any defects exist. For example, in this embodiment, monitoring the warp-knitted upper production line from 8:00 to 8:15 on a certain day yielded 274 entries, including 102 pink tri-twill and 172 blue plaid. On the manufacturer's backend monitoring monitor, each warp-knitted jacquard upper flowing through the shoe production line had a marker box displayed in real time above and behind the upper on the page. Within the marker box, the monitoring time, upper fabric type and its ranking within the fabric category, product category of the upper and its ranking within that category, upper color and its ranking within the color category, and information on whether defects exist and their severity are displayed.

[0059] In addition, this embodiment also includes determining the statistical results based on the photoelectric effect;

[0060] The photoelectric effect is used to detect whether warp-knitted jacquard shoe uppers are passing through the production line, and a trigger count is performed based on the detection results; in this embodiment, a photoelectric detector is used to detect whether warp-knitted jacquard shoe uppers are passing through the production line.

[0061] The trigger count result is compared with the digital tag result to obtain the judgment digital tag result. It should be noted that in this embodiment, the judgment of the quantity statistics result based on the photoelectric effect is not continuous, but rather determined according to the needs of the actual application; for example, sampling activation, where the photodetector is activated during a selected time period for mutual verification between the dual quantity statistics method and the quantity statistics method based on monitoring images. This further improves the intelligence level, work efficiency, and accuracy of warp-knitted shoe upper processing.

[0062] The range of three-dimensional parameters of the finite element model is corrected based on material and industrial errors, specifically including:

[0063] To obtain the maximum elasticity of different knitted materials;

[0064] The material influence weights of the three-dimensional parameters of the finite model are calculated based on the maximum elasticity of different knitted materials.

[0065] Using real-time knitted materials as the object, the deformation data of warp-knitted jacquard shoe uppers in different directions during circumferential stretching were extracted by measuring the maximum deformation of the warp-knitted jacquard shoe uppers.

[0066] The range of three-dimensional parameters of the finite element model is determined by using material influence weights and deformation data, thus completing the correction of the three-dimensional parameter range of the finite element model.

[0067] Preferably, the target detection network structure is as follows:

[0068] Input layer - first convolutional layer - feature extraction layer - fully connected layer - second convolutional layer - feature extraction layer - fusion layer - output layer.

[0069] The process of obtaining information regarding defects in the content of digital tags specifically includes:

[0070] In the object detection network, the first convolutional layer performs convolution processing on the image of flawless warp-knitted shoe uppers, constructs the fitness function using the Fisher criterion, and calculates the function value corresponding to each particle;

[0071] The quantum behavior particle swarm optimization algorithm is used to optimize the filtering parameters in the target detection network to obtain the optimal filtering parameters.

[0072] Using the optimal filtering parameters, the second convolutional layer performs convolution processing on the framed images, and then performs binarization processing to obtain the defect information in the framed images.

[0073] The fitness function expression is:

[0074] Where f(x) represents the function value corresponding to each particle; β is the constant coefficient of the function representing different image sizes; h(x) represents the mean energy of the flawless warp-knitted shoe upper image after convolution processing; g(y) represents the standard deviation of the flawless warp-knitted shoe upper image after convolution processing.

[0075] Example 2:

[0076] This embodiment discloses a terminal device, which includes a memory, a processor, and a warp-knitted jacquard shoe upper quantity statistics and quality inspection program stored in the memory and executable on the processor. When the processor executes the warp-knitted jacquard shoe upper quantity statistics and quality inspection program, it implements the steps of a warp-knitted jacquard shoe upper quantity statistics and quality inspection method.

[0077] The methods and apparatus disclosed in the embodiments are described simply because they correspond to the content disclosed in the embodiments; relevant details can be found in the description.

[0078] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for quantity statistics and quality inspection of warp-knitted jacquard shoe uppers, characterized in that, include: Video monitoring was collected at the finishing station of warp-knitted jacquard shoe uppers on the shoe manufacturing production line. Based on target detection technology and video processing technology, the warp-knitted jacquard shoe uppers in the monitoring video are identified and digitally marked; Specifically, it includes: Obtain a training sample set for the detection of warp-knitted jacquard shoe uppers; establish a target detection network, and train the target detection network using the training sample set until the target detection results reach a predetermined standard; The monitoring video is processed into frames; the frame images of a continuous fixed time period are used as network inputs and input into the target detection network, and the recognition result of the warp-knitted jacquard shoe upper is output; Calculate the influence weights of shoe surface curvature and light intensity on the monitored video pixels; and use the influence weights to optimize the recognition results. The optimized identification results are sorted according to monitoring time, and the sorting order is marked with numbers to obtain the content of the number markings; The digital markers include: monitoring time, which warp-knitted jacquard upper, its category information and its sorting position within that category, and whether there are any defects. A finite element model of the shoe upper was established based on the standard warp-knitted jacquard shoe upper; The range of three-dimensional parameters in the finite element model is corrected based on material and industrial errors; specifically including: To obtain the maximum elasticity of different knitted materials; The material influence weights of the three-dimensional parameters of the finite element model are calculated based on the maximum elasticity of the different knitted materials. Using real-time knitted materials as the object, the deformation data of the warp-knitted jacquard shoe upper in different directions during circumferential stretching were extracted by measuring the maximum deformation of the warp-knitted jacquard shoe upper. The range of the three-dimensional parameters of the finite element model is determined by using the material influence weights and the deformation data, thereby completing the correction of the range of the three-dimensional parameters of the finite element model. Using scanning principles, a three-dimensional scan of the warp-knitted jacquard shoe upper to be tested was performed to obtain three-dimensional parameters; The three-dimensional parameters are matched and determined with the range of the three-dimensional parameters of the corrected finite element model; The content of the digital marker is updated based on the matching result.

2. The method for quantity counting and quality inspection of warp-knitted jacquard shoe uppers according to claim 1, characterized in that, It also includes determining the statistical results based on the photoelectric effect; The photoelectric effect is used to detect whether the warp-knitted jacquard shoe upper passes through the production line, and trigger counting is performed based on the detection results; The trigger count result is compared with the digital tag result to obtain the determination digital tag result.

3. The method for quantity counting and quality inspection of warp-knitted jacquard shoe uppers according to claim 1, characterized in that, The target detection network structure is as follows: Input layer - first convolutional layer - feature extraction layer - fully connected layer - second convolutional layer - feature extraction layer - fusion layer - output layer.

4. The method for quantity counting and quality inspection of warp-knitted jacquard shoe uppers according to claim 1, characterized in that, The process of obtaining information on whether there are defects in the content of the digital markers specifically includes: In the target detection network, the first convolutional layer performs convolution processing on the image of a flawless warp-knitted shoe upper, constructs a fitness function using the Fisher criterion, and calculates the function value corresponding to each particle; The filtering parameters in the target detection network are optimized using the quantum behavior particle swarm optimization algorithm to obtain the optimal filtering parameters. Using the optimal filtering parameters, the second convolutional layer performs convolution processing on the framed images, and then performs binarization processing to obtain the defect information in the framed images.

5. The method for quantity counting and quality inspection of warp-knitted jacquard shoe uppers according to claim 4, characterized in that, The fitness function expression is: ; in This represents the function value corresponding to each particle; These are the constant coefficients of the function representing different image sizes; This represents the average energy of the flawless warp-knitted shoe upper image after convolution processing; This represents the standard deviation of the image of the flawless warp-knitted upper after convolution processing.

6. A terminal device, characterized in that, The terminal device includes a memory, a processor, and a warp-knitted jacquard shoe upper quantity counting and quality inspection program stored in the memory and executable on the processor. When the processor executes the warp-knitted jacquard shoe upper quantity counting and quality inspection program, it implements the steps of the warp-knitted jacquard shoe upper quantity counting and quality inspection method as described in any one of claims 1-5.

Images