Textile fabric pattern recognizing and cutting system

Abstract

The invention discloses a textile fabric pattern recognizing and cutting system, and belongs to the field of textile fabric, the recognizing and cutting system comprises a multi-station cutting platform and an intelligent control system, the multi-station cutting platform comprises a cutting machine tool, a control system module, an intelligent pattern recognizing system module and a transmission system module, the intelligent pattern recognition system module is used for recognizing and positioning marks or patterns on the materials; through cooperative use of all the devices and arrangement of an intelligent pattern recognition system module, specific feature information analysis of different pieces of cloth is achieved, so that the specific cutting technology needing to be adopted by the cloth is rapidly obtained, and the cutting efficiency of the cloth is improved through combined statistics of color features, texture features and material features of the cloth. And feature recognition is completed based on the intelligent model recognition module, and the intelligent model recognition module is used for analyzing the recognition result to perform matching of process selection, so that the most adaptive cutting pattern is completed, and the cloth is prevented from adopting an unmatched cutting process.

Description

Technical Field

[0001] The present invention relates to the technical field of textile fabrics, and particularly to a textile fabric pattern recognition and cutting system. Background Art

[0002] Fabrics are commonly used materials in decorative materials, including various fabrics such as chemical fiber carpets, non-woven wall coverings, linen, nylon, colored adhesive tapes, flannel, etc. Fabrics play a quite important role in decorative displays and are often the main force that cannot be ignored in the entire sales space. A large amount of fabrics are used for wall facades, partitions, and background treatments, which can also form a good commercial space display style. Textile fabrics are indispensable materials in modern life, covering a wide range of types and uses. Textile fabrics are materials processed from various fibers through different textile processes and are widely used in multiple fields such as clothing, home, and industry. According to the type of fiber, weaving method, and use, fabrics can be divided into many different types. Textile fabric pattern recognition is an important technology that has a positive effect on improving production efficiency and product quality. With the progress of computer vision and artificial intelligence technologies, the accuracy and application scope of pattern recognition are continuously expanding. Pattern recognition is an important topic in the fields of computer vision and artificial intelligence, which involves identifying and classifying specific patterns, shapes, or features from images. A pattern recognition and cutting system is a system that uses computer vision technology, combined with cutting equipment (such as laser cutting, knife cutting, etc.), to perform precise cutting by automatically identifying and processing patterns on fabrics. This system can improve the efficiency, accuracy, and flexibility of cutting and is widely used in industries such as clothing manufacturing, textile production, furniture, and automotive interiors.

[0003] In the existing fabric pattern recognition and cutting system, in pattern recognition and analysis, if the selected cutting process is not suitable for the characteristics of the fabric, it will lead to a decline in cutting quality, such as frayed edges, tearing, or inaccurate cutting lines. The wrong cutting process may cause fabric waste. Since the best cutting method cannot be determined in advance, it may result in waste of raw materials, thereby significantly increasing production costs. It is impossible to specifically identify and analyze different fabrics and obtain the cutting process required for the fabric. There is a lack of automated pattern recognition and cutting process planning, which may require more manual intervention, increasing labor costs and the possibility of errors.

[0004] Therefore, the present invention provides a textile fabric pattern recognition and cutting system to solve the above problems. Summary of the Invention

[0005] (1) Technical Problems to be Solved

[0006] The present invention provides a textile fabric pattern recognition and cutting system, aiming to solve the problems raised in the background art.

[0007] (2) Technical Solution

[0008] To achieve the above object, the present invention provides the following technical solution: The identification and cutting system consists of a multi-station cutting platform and an intelligent control system;

[0009] The multi-station cutting platform includes a cutting machine tool, a control system module, an intelligent pattern recognition system module, and a transmission system module. The intelligent pattern recognition system module is used to identify and locate marks or patterns on the material;

[0010] The intelligent pattern recognition system module includes an image acquisition module, a model intelligent recognition module, and an identification result classification and process selection module. The image acquisition module is used to collect fabric images and feature extraction. The model intelligent recognition module completes feature recognition based on the established recognition model. The identification result classification and process selection module is used to match the process selection through the analysis of the completed identification results.

[0011] As a preferred technical solution of this application, the image acquisition module combines and statistically analyzes the color features, texture features, and material features of the fabric, and conducts model analysis based on the above features.

[0012] As a preferred technical solution of this application, the image acquisition module converts the image from the RGB color space to a color space that is more in line with human eye perception.

[0013] As a preferred technical solution of this application, the image acquisition module uses the method of Gabor filter to achieve texture feature acquisition. The Gabor filter is a linear filter. The texture feature is localized based on the Gaussian function and the sine or cosine function. The image acquisition module uses the cv2.getGaborKernel() function to generate a Gabor filter and performs a filtering operation on the fabric image to extract the filtered texture features;

[0014] The image acquisition module uses the method of convolutional neural network to complete the acquisition of fabric material features. The convolutional neural network can automatically learn fabric features and process large-scale data. By converting the fabric texture into an image and inputting it into the convolutional neural network, and performing in-depth processing based on the ResNet network structure, the fabric material features can be extracted. The output results of the image acquisition module include cotton, linen, silk, or polyester fiber.

[0015] As a preferred technical solution of this application, the identification result classification and process selection module conducts specific analysis based on the above-acquired fabric features, and uses the machine learning algorithm of support vector machine for training to match the best cutting process for the fabric to be cut.

[0016] As a preferred technical solution of the present application, the recognition model in the intelligent pattern recognition system module specifically includes: for the recoverable area of the image texture in the collected fabric pattern, through frequency extraction is performed; and based on the extracted frequency, it is judged whether it is a continuous frequency domain or a discrete frequency domain, and energy calculation is completed based on the obtained frequency domain characteristics, and finally Gabor filtering is obtained;

[0017] Among them, Gabor filtering is through

[0018]

[0019]

[0020] extract multi-scale and multi-directional texture information, and perform autonomous recognition according to the obtained texture information.

[0021] As a preferred technical solution of the present application, the recognition result classification and process selection module further includes using historical cutting process data as the label data of the model, and combining the extracted fabric features and the corresponding cutting processes for multiple model trainings, so as to learn the relationship between the features and the cutting processes through the machine learning algorithm of the support vector machine, and use it to complete the feature recognition and the matching of specific cutting processes. The fabric cutting processes include laser cutting, tool cutting, ultrasonic cutting and waterjet cutting. Laser cutting is used for thinner synthetic fiber fabrics, tool cutting is used for thick materials with higher density, ultrasonic cutting is used for soft and easily frayed materials, and waterjet cutting is used for multi-layer fabrics and natural fiber fabrics.

[0022] As a preferred technical solution of the present application, through the characteristics of the fabric itself, such as texture, material and color, the texture is used as parameter α, the material is used as parameter β, and the color is used as parameter θ, a matching model with α, β and θ as variables is established, and y is used as the output result. The formula is established as follows: y = α T β + θ. After collection and analysis, the specific characteristics of the fabric are obtained, and the specific cutting process is completed according to the characteristics.

[0023] (III) Beneficial effects

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows: By setting up an intelligent pattern recognition system module, the specific characteristic information of different fabrics can be analyzed, so as to quickly obtain the specific cutting process required for the fabric. Through the combination and statistics of the color characteristics, texture characteristics and material characteristics of the fabric, and based on the model intelligent recognition module to complete the feature recognition, the matching of process selection is carried out for the analysis of the recognition results, and the most suitable cutting pattern is completed, avoiding the use of mismatched cutting processes for the fabric. Incorrect cutting processes will lead to fabric waste. The use of the intelligent pattern recognition system module can realize intelligent and high-precision textile fabric cutting operations, effectively improve the production efficiency and product quality of textile fabric processing, reduce labor costs and resource waste, improve the product quality of the finally completed fabric, and the applicable cutting method can effectively reduce the risks of defects and losses caused by improper processes, ensuring stable production. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic diagram of the steps of a textile fabric pattern recognition and cutting system;

[0026] Figure 2 It is a schematic diagram of the process of the image acquisition module in a textile fabric pattern recognition and cutting system. DETAILED DESCRIPTION OF THE INVENTION

[0027] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

[0028] The present invention provides a textile fabric pattern recognition and cutting system, as Figure 1 - Figure 2 shown. The textile fabric pattern recognition and cutting system includes a multi-station cutting platform and an intelligent control system.

[0029] The multi-station cutting platform includes a cutting machine tool, a control system module, an intelligent pattern recognition system module and a transmission system module. The intelligent pattern recognition system module is used to identify and locate the marks or patterns on the material.

[0030] The intelligent pattern recognition system module includes an image acquisition module, a model intelligent recognition module, and a recognition result classification and process selection module. The image acquisition module is used to collect fabric images and extract features. The model intelligent recognition module completes feature recognition based on the established recognition model. The recognition result classification and process selection module is used to match the process selection for the analysis of the recognition results.

[0031] Through the set intelligent pattern recognition system module, the specific characteristic information of different fabrics can be analyzed, so as to quickly obtain the specific cutting process required for the fabric. By combining and statistically analyzing the color characteristics, texture characteristics and material characteristics of the fabric, and based on the model intelligent recognition module to complete the feature recognition, the matching of process selection is carried out for the results of the recognition, and the most suitable cutting pattern is completed, avoiding the use of mismatched cutting processes for the fabric. Incorrect cutting processes will lead to fabric waste. The use of the intelligent pattern recognition system module can realize intelligent and high-precision textile fabric cutting operations, effectively improve the production efficiency and product quality of textile fabric processing, reduce labor costs and resource waste, and improve the product quality of the finally completed fabric.

[0032] The image acquisition module combines and statistically analyzes the color characteristics, texture characteristics and material characteristics of the fabric, and conducts model analysis based on the above characteristics.

[0033] The image acquisition module converts the image from the RGB color space to a color space that is more in line with human eye perception (such as the HSV color space). The formula for converting RGB to the HSV color space is as follows:

[0034] V = max(R, G, B)

[0035] Δ = V - min(R, G, B)

[0036] S = 0 (if V == 0) else (Δ / V)

[0037] H = (if Δ == 0) 0

[0038] (if V == R) (60 * ((G - B) / Δ) + 360) % 360

[0039] (if V == G) (60 * ((B - R) / Δ) + 120) % 360

[0040] (if V == B) (60 * ((R - G) / Δ) + 240) % 360.

[0041] Based on the above formula, the color information of the image can be intuitively processed and analyzed to extract the specific color of the fabric. By using the conversion of the color space, a color space that is more in line with human eye perception can be obtained. By selecting a suitable color space, specific image processing tasks can be carried out more effectively, which helps to improve the accuracy of fabric color feature extraction. Among them, V represents brightness, S represents saturation, H represents hue, and Δ represents chroma.

[0042] The image acquisition module uses the Gabor filter method to implement texture feature acquisition. The Gabor filter is a linear filter. The texture features are localized based on the Gaussian function and the sine or cosine function. The image acquisition module uses the cv2.getGaborKernel() function to generate the Gabor filter and perform a filtering operation on the cloth image to extract the filtered texture features. Using the Gabor filter, multi-scale and multi-directional texture information can be extracted from the cloth image. The Gabor function can be expressed as:

[0043]

[0044]

[0045] The above formula consists of the continuous frequency domain Gabor wavelet filter (g r ) and the discrete frequency domain Gabor wavelet filter (g i ), which are as follows:

[0046]

[0047]

[0048]

[0049]

[0050] x0 and y0 are the center frequency points, and a and b describe the shape of the filtering window. The partial code of the program is as follows:

[0051] Function implementation: Create a Gabor filter

[0052]

[0053] After filtering, the extraction of the image texture is finally obtained.

[0054] Where x and y are two-dimensional coordinates, and g i is the discrete frequency domain Gabor wavelet filter.

[0055] The image acquisition module uses the method of convolutional neural network to complete the acquisition of fabric material characteristics. The convolutional neural network can automatically learn fabric characteristics and process large-scale data. By converting the fabric texture into an image and inputting it into the convolutional neural network, and performing deep processing based on the ResNet network structure, the fabric material characteristics can be extracted. The output results of the image acquisition module include cotton, linen, silk or polyester fiber. By obtaining the specific fabric material, the cutting process required for the fabric can be finally determined, which is convenient for realizing intelligent and high-precision textile fabric cutting operations, effectively improving the production efficiency and product quality of textile fabric processing, reducing labor costs and resource waste, and improving the product quality of the finally completed fabric.

[0056] The recognition result classification and process selection module conducts specific analysis based on the fabric characteristics collected above, uses the machine learning algorithm of support vector machine for training, and matches the best cutting process for the fabric to be cut. After multiple cutting trainings, the production efficiency can be improved, material waste can be reduced, and the cutting quality can be ensured to meet the requirements of the final finished product. It can calculate the cutting requirements that need to be used based on the fabric characteristics, which is convenient for improving the fabric cutting quality.

[0057] In the intelligent pattern recognition system module, the recognition model specifically includes: for the recoverable area of the image texture in the collected fabric pattern, frequency extraction is performed; and based on the extracted frequency, it is judged whether it is a continuous frequency domain or a discrete frequency domain, and the energy calculation is completed based on the obtained frequency domain characteristics, and finally the Gabor filter is obtained;

[0058] Among them, the Gabor filter

[0059]

[0060]

[0061] extracts multi-scale and multi-directional texture information, and performs autonomous recognition according to the obtained texture information.

[0062] The recognition result classification and process selection module also includes using the historical cutting process data as the label data of the model, and combining the extracted fabric characteristics and the corresponding cutting processes for multiple model trainings, so as to learn the relationship between the characteristics and the cutting processes through the machine learning algorithm of support vector machine, and be used to complete feature recognition and the matching of specific cutting processes. The fabric cutting processes include laser cutting, tool cutting, ultrasonic cutting and water jet cutting. The laser cutting is used for thinner synthetic fiber fabrics, the tool cutting is used for thicker materials with higher density, the ultrasonic cutting is used for soft and easily frayed materials, and the water jet cutting is used for multi-layer fabrics and natural fiber fabrics.

[0063] Based on the characteristics of the fabric itself, such as texture, material, and color, with texture as parameter α, material as parameter β, and color as parameter θ, a matching model with α, β, and θ as variables is established, and y is the output result. The formula is as follows: y = α T β + θ. Through collection and analysis, the specific characteristics of the fabric are obtained, and the specific cutting process is completed according to the characteristics.

[0064] In the specific implementation, when identifying the fabric, the image acquisition module is used to collect the color, texture, and material characteristics of the fabric. The following formula is used to convert the image from the RGB color space to the HSV color space.

[0065] V = max(R, G, B)

[0066] Δ = V - min(R, G, B)

[0067] S = 0 (if V == 0) else (Δ / V)

[0068] H = (if Δ == 0) 0

[0069] (if V == R)(60 * ((G - B) / Δ) + 360) % 360

[0070] (if V == G)(60 * ((B - R) / Δ) + 120) % 360

[0071] (if V == B)(60 * ((R - G) / Δ) + 240) % 360, where V represents brightness, S represents saturation, H represents hue, and Δ represents tone. With this formula, the color information of the image can be intuitively processed and analyzed to extract the specific color of the fabric. Subsequently, the Gabor filter is used by the image acquisition module to achieve texture acquisition. Localization is achieved through the Gaussian function and the sine or cosine function. Then, the cv2.getGaborKernel() function is used to generate the Gabor filter to perform filtering operations on the fabric image, thereby extracting the filtered texture features. Subsequently, the method of convolutional neural network is used to acquire the material features of the fabric. By converting the fabric texture into an image and inputting it into the convolutional neural network, deep processing is performed based on the ResNet network structure to extract the fabric material features, and it is concluded that the fabric is cotton, linen, silk, or polyester fiber. Specifically, which category of fabric it is. Subsequently, after the specific features of the fabric are completed, specific analysis is carried out. Using the machine learning algorithm of support vector machine, with the historical cutting process data as the base data, different cutting processes are suitable for different types of fabrics. Through reasonable selection, the accuracy and effect of cutting can be ensured, errors can be reduced. Combining the extracted fabric features and the corresponding fabric cutting processes for multiple models, comparing the compatibility between the fabric features and the cutting processes, and performing the final analysis to complete the feature recognition and the matching of the specific cutting process, to match the most suitable cutting process and cutting pattern for the textile fabric. According to the fabric material, the fabric is placed on the workbench. There is a nozzle at the top of the fabric, and it is sprayed on the fabric according to the required number and size of the patterns. The system is used to identify each pattern and cut according to the size of the pattern. After cutting one layer of the fabric, the fabric is rolled away by the reel to facilitate cutting the next layer of the fabric, which is convenient for realizing intelligent and high-precision textile fabric cutting operations, effectively improving the production efficiency and product quality of textile fabric processing, reducing labor costs and resource waste, and improving the product quality of the finally completed fabric. The applicable cutting method can effectively reduce the risks of defects and losses caused by improper processes and ensure stable production.

[0072] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A textile fabric pattern recognition and cutting system, characterized in that: The described identification and cutting system includes a multi-station cutting platform and an intelligent control system; The multi-station cutting platform includes a control system module, an intelligent pattern recognition system module, and a transmission system module. The intelligent pattern recognition system module is used to identify and locate marks or patterns on the material; The intelligent pattern recognition system module includes an image acquisition module, a model intelligent recognition module, and an identification result classification and process selection module. The image acquisition module is used to collect fabric images and extract features. The model intelligent recognition module completes feature recognition based on the established recognition model. The identification result classification and process selection module is used for the analysis of the obtained results and the matching of process selection.

2. The textile fabric pattern recognition and cutting system according to claim 1, characterized in that: The image acquisition module combines statistics by collecting the color features, texture features, and material features of the fabric, and performs model analysis based on the above features.

3. The textile fabric pattern recognition and cutting system according to claim 1, characterized in that: The image acquisition module converts the image from the RGB color space to a color space that conforms to human eye perception.

4. A textile fabric pattern recognition and cutting system according to claim 1, characterized in that: The image acquisition module uses the method of Gabor filter to achieve texture feature acquisition. The Gabor filter is a linear filter. The texture feature is localized based on the Gaussian function and the sine or cosine function. The image acquisition module uses the cv2.getGaborKernel() function to generate the Gabor filter and perform filtering operations on the fabric image to extract the filtered texture features; The image acquisition module uses the method of convolutional neural network to complete the acquisition of fabric material features. The convolutional neural network can automatically learn fabric features and process large-scale data. By converting the fabric texture into an image and inputting it into the convolutional neural network, and performing deep processing based on the ResNet network structure, the fabric material features can be extracted. The output results of the image acquisition module include cotton, linen, silk, or polyester fiber.

5. The textile fabric pattern recognition and cutting system according to claim 1, characterized in that: The identification result classification and process selection module performs specific analysis based on the fabric features collected above, and uses the machine learning algorithm of support vector machine for training to match the best cutting process for the fabric to be cut.

6. The textile fabric pattern recognition and cutting system according to claim 1, wherein: In the recognition model of the intelligent pattern recognition system module, specifically, for the recoverable area of the image texture in the collected fabric pattern, through frequency extraction is performed; and based on the extracted frequency, it is judged whether it is a continuous frequency domain or a discrete frequency domain, and based on the obtained frequency domain characteristics, energy calculation is completed, and finally Gabor filtering is obtained; Among them, Gabor filtering is carried out through extracting multi-scale and multi-directional texture information and performing autonomous recognition according to the obtained texture information.

7. A textile fabric pattern recognition and cutting system according to claim 1, characterized in that: The identification result classification and process selection module also includes using historical cutting process data as the label data of the model, combining the extracted fabric features and the corresponding cutting processes for multiple model trainings, so as to learn the relationship between features and cutting processes through the machine learning algorithm of support vector machine, and be used to complete feature recognition and the matching of specific cutting processes. The fabric cutting processes include laser cutting, knife cutting, ultrasonic cutting, and water jet cutting. Laser cutting is used for thinner synthetic fiber fabrics. Knife cutting is used for thick materials with higher density. Ultrasonic cutting is used for soft and easily frayed materials. Water jet cutting is used for multi-layer fabrics and natural fiber fabrics.

8. A textile fabric pattern recognition and cutting system according to claim 1, characterized in that: Through the characteristics of the fabric itself, such as texture, material, and color, with texture as parameter α, material as parameter β, and color as parameter θ, a matching model with α, β, and θ as variables is established, and y is used as the output result. The formula is established as follows: y = αTβ + θ. Through collection and analysis, the specific characteristics of the fabric are obtained, and the specific cutting process is completed according to the characteristics.

Images