Cloth spreading machine fabric defect online detection method

Abstract

The invention discloses an online detection method for fabric defects of a cloth paving machine. The online detection method comprises the following steps: 1) collecting surface image information of a fabric in real time; 2) image preprocessing: preprocessing the surface image information; 3) according to the fabric type, calling a corresponding defect detection algorithm to carry out defect identification on the surface image information, and obtaining the category and position information of defects; 4) storing category and position information of the defects; and the category and position information of the defects is transmitted to a display screen for visual display. According to the online detection method, the defects in the cloth paving process of the cloth paving machine are detected in real time, defect information is accurately obtained, the detection efficiency is high, and the product quality is ensured.

Description

Technical Field

[0001] The invention relates to the technical field of fabric defect detection, and in particular to an online fabric defect detection method for a spreading machine. Background Art

[0002] In the process of clothing production, fabrics will inevitably have defects, such as holes or damage. If these defects appear on the cut pieces, it will directly lead to the scrapping of the pieces, which will affect the subsequent production process and the overall production rhythm. At present, the detection of fabric defects mainly relies on the naked eye observation of manual operators, which is not only inefficient, but also has the risk of missing defects, which may lead to quality problems in the final product. In order to overcome the above problems, it is urgent to provide an online detection method for fabric defects of a spreading machine, which can detect the quality of fabrics in real time during the automatic spreading process, timely discover and record the location of defects, and help staff quickly lock the layer number and actual location of the waste piece, thereby improving the efficiency of fabric defect detection and promoting the improvement of product quality and production efficiency. Summary of the invention

[0003] In view of the above-mentioned defects or deficiencies in the prior art, it is desired to provide an online method for detecting fabric defects of a spreading machine, which realizes real-time detection of defects during the spreading process of the spreading machine, accurately obtains defect information, has high detection efficiency, and ensures product quality.

[0004] The present invention provides an online method for detecting fabric defects of a spreading machine, comprising the following steps:

[0005] 1) Real-time collection of surface image information of fabrics during the working process of the spreading machine;

[0006] 2) Image preprocessing, preprocessing the surface image information to reduce high-frequency noise and salt and pepper noise;

[0007] 3) According to the fabric type, the corresponding defect detection algorithm is called to identify the defects on the preprocessed surface image information to obtain the category and location information of the defects; the defect detection algorithm includes knitted fabric defect detection, woven fabric defect detection and stretch fabric defect detection; the location information of the defect includes the number of fabric layers N where the defect is located and the physical coordinates (X w ,Y w );

[0008]

[0009] Among them, the formula Indicates rounding up;

[0010] v is the spreading speed of the spreading machine, in m / min;

[0011] T1 is the time when the defect is detected, T0 is the time when the cloth is started, both in min;

[0012] l is the single-layer spreading length of the spreading machine, in meters;

[0013]

[0014]

[0015]

[0016]

[0017]

[0018] Among them, (u,v) is the center coordinate of the defect;

[0019] (min_x, min_y) is the coordinate of the upper left corner of the defect boundary box, and (max_x, max_y) is the coordinate of the lower right corner of the defect boundary box;

[0020] M is the homography matrix;

[0021] 4) Store the category and location information of the defects; and transmit the category and location information of the defects to the display screen for visual display.

[0022] Furthermore, the surface image information is smoothed by a mean filter and a median filter respectively to reduce high frequency noise and salt and pepper noise.

[0023] Furthermore, the knitted fabric defect detection comprises the following steps:

[0024] 311) Enhance the contrast of surface image information;

[0025] 312) Use texture analysis algorithm to detect needle leakage defects, needle skipping defects and pilling defects;

[0026] 313) Use edge detection algorithm to identify and locate small needle leakage defects or needle skipping defects.

[0027] Furthermore, the woven fabric defect detection comprises the following steps:

[0028] 321) Use frequency domain analysis to detect periodic defects;

[0029] 322) Establish a normal fabric template and use template matching technology to detect and locate defective areas that deviate from the template;

[0030] 323) In view of the color difference and uneven thickness of fabrics, the color and brightness anomalies are identified through color space transformation and brightness histogram analysis.

[0031] Furthermore, the elastic fabric defect detection comprises the following steps:

[0032] 331) Use dynamic detection technology to detect defects in stretch fabrics and evaluate the changes in defects in stretch fabrics under stretching;

[0033] 332) Use multi-scale analysis methods to detect defects at different scales.

[0034] Furthermore, in step 3), when a defect is detected, a contour extraction algorithm is used to perform statistical analysis on the size and shape of the defect in the surface image information, and each pixel point (x, y) in the defect area is traversed to obtain the upper left corner coordinates (min_x, min_y) and lower right corner coordinates (max_x, max_y) of the defect boundary box; and then the center coordinates (u, v) of the defect are calculated; and then the pixel coordinates of the defect are converted into actual physical coordinates (X, Y) on the fabric surface according to the homography matrix M of the image acquisition device. w ,Y w ).

[0035] Furthermore, the stored category and location information of the defect is transferred and used for neural network learning to improve the robustness of the defect detection algorithm.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] The detection method of the present invention realizes the real-time detection of fabric defects in the production process, accurately obtains defect information, has high detection efficiency, avoids bad fabrics from entering subsequent processes, and ensures the quality of the product.

[0038] By storing and analyzing test data, we can continuously improve defect detection algorithms, improve the accuracy of defect detection, and help garment factories develop in the direction of intelligence and refinement, which has extremely high promotion value.

[0039] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0041] Figure 1 The figure is a flow chart of the online detection method for fabric defects on a spreading machine. DETAILED DESCRIPTION

[0042] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It should also be noted that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0043] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0044] Please refer to Figure 1 The embodiment of the present invention provides an online detection method for fabric defects of a spreading machine, comprising the following steps:

[0045] 1) The surface image information of the fabric during the working process of the spreading machine is collected in real time through a high-definition camera installed above the spreading machine;

[0046] 2) Image preprocessing: smoothing the surface image information through mean filter and median filter to reduce high-frequency noise and salt and pepper noise;

[0047] 3) According to the fabric type, the corresponding defect detection algorithm is called to identify defects on the preprocessed surface image information to obtain the category and location information of the defects;

[0048] Defect detection algorithms include knitted fabric defect detection, woven fabric defect detection, and stretch fabric defect detection;

[0049] Defects in knitted fabrics usually manifest as discontinuities or anomalies in the texture. A special knitted fabric defect detection method is designed to detect these structural features, including the following steps:

[0050] 311) Enhance the contrast of surface image information, improve the visibility of defects, and make it easier for detection algorithms to identify irregular texture changes;

[0051] 312) Use texture analysis algorithms, such as Gabor filter, local binary pattern (LBP), etc., to detect needle leakage defects, needle skipping defects and pilling defects;

[0052] 313) Use Canny, Sobel and other edge detection algorithms to identify and locate small needle leakage defects or needle skipping defects;

[0053] Woven fabrics have a regular weaving structure. According to this structural feature, a special woven fabric defect detection method is designed, which includes the following steps:

[0054] 321) Use frequency domain analysis (such as Fourier transform) to detect periodic defects such as weft yarn breaks and yarn defects;

[0055] 322) Establish a normal fabric template and use template matching techniques (such as cross-correlation and structural similarity index) to detect and locate defective areas that deviate from the template;

[0056] 323) For the color difference and uneven thickness of fabrics, the color and brightness anomalies are identified through color space transformation (such as RGB to HSV) and brightness histogram analysis;

[0057] Stretch fabric refers to fabric woven with stretch yarn or special structure, which is easy to deform during stretching. According to the structural characteristics, a special stretch fabric defect detection is designed, including the following steps:

[0058] 331) Use dynamic detection technology (such as motion estimation based on optical flow method) to detect defects of elastic fabric and evaluate the change of defects of elastic fabric under stretch state;

[0059] 332) Since the defects of stretch fabric may appear differently at different scales, multi-scale analysis methods (such as wavelet transform) are used to detect defects at different scales;

[0060] The location information of the defect includes the number of the ply N where the defect is located and the physical coordinates (X w ,Y w ); when a defect is detected, the size and shape of the defect in the surface image information are statistically analyzed using a contour extraction algorithm. By traversing each pixel point (x, y) in the defect area, the upper left corner coordinates (min_x, min_y) and lower right corner coordinates (max_x, max_y) of the defect boundary box are obtained; then the center coordinates (u, v) of the defect are calculated; and then the pixel coordinates of the defect are converted into actual physical coordinates (X w ,Y w ).

[0061]

[0062] Among them, the formula Indicates rounding up;

[0063] v is the spreading speed of the spreading machine, in m / min;

[0064] T1 is the time when the defect is detected, T0 is the time when the cloth is started, both in min;

[0065] l is the single-layer spreading length of the spreading machine, in meters;

[0066]

[0067]

[0068]

[0069]

[0070]

[0071] Among them, (u,v) is the center coordinate of the defect;

[0072] (min_x, min_y) is the coordinate of the upper left corner of the defect boundary box, and (max_x, max_y) is the coordinate of the lower right corner of the defect boundary box;

[0073] M is the homography matrix;

[0074] 4) Store the category and location information of the defects; and transmit the category and location information of the defects to the display screen for visual display; transfer the stored category and location information of the defects for neural network learning to improve the robustness of the defect detection algorithm.

[0075] In this embodiment, the detection system for the online detection method provides a data storage module and a data analysis module, which are integrated in the computer to store the defect information data of the fabric in real time, classify and manage it, and ensure the integrity and security of the data. The detection system performs statistical analysis on the defect information data, and transmits the defect information on each layer of fabric to the display screen for visual display. Workers can optimize the arrangement order of paper patterns according to the distribution of defects, avoid defect locations, and avoid or reduce fabric waste in the cutting process as much as possible.

[0076] The stored defect categories and location information can be used to train various defect detection algorithms. The defect detection algorithm model (including but not limited to convolutional neural network (CNN) model based on deep learning, support vector machine (SVM), etc.) is used to learn the defect data of different types of fabrics, improve the recognition accuracy and robustness of each defect detection algorithm, and thus continuously improve the reliability of defect detection.

[0077] In the description of this specification, the description of the terms "one embodiment", "some embodiments", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0078] The above are only preferred embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for online detection of fabric defects on a spreading machine, characterized in that: The steps include: 1) Real-time collection of surface image information of fabrics during the working process of the spreading machine; 2) Image preprocessing, preprocessing the surface image information to reduce high-frequency noise and salt and pepper noise; 3) According to the fabric type, the corresponding defect detection algorithm is called to identify the defects on the preprocessed surface image information to obtain the category and location information of the defects; the defect detection algorithm includes knitted fabric defect detection, woven fabric defect detection and stretch fabric defect detection; the location information of the defect includes the number of fabric layers N where the defect is located and the physical coordinates (X w ,Y w ); Among them, the formula Indicates rounding up; v is the spreading speed of the spreading machine, in m / min; T1 is the time when the defect is detected, T0 is the time when the cloth is started, both in min; l is the single-layer spreading length of the spreading machine, in meters; Where (u,v) is the center coordinate of the defect; (min_x, min_y) is the coordinate of the upper left corner of the defect boundary box, and (max_x, max_y) is the coordinate of the lower right corner of the defect boundary box; M is the homography matrix; 4) Store the category and location information of the defects; and transmit the category and location information of the defects to the display screen for visual display.

2. The method for online detection of fabric defects on a spreading machine according to claim 1, characterized in that: In the step 2), the surface image information is smoothed by a mean filter and a median filter respectively to reduce high-frequency noise and salt and pepper noise.

3. The method for online detection of fabric defects on a spreading machine according to claim 1, characterized in that: The knitted fabric defect detection comprises the following steps: 311) Enhance the contrast of surface image information; 312) Use texture analysis algorithm to detect needle leakage defects, needle skipping defects and pilling defects; 313) Use edge detection algorithm to identify and locate small needle leakage defects or needle skipping defects.

4. The method for online detection of fabric defects on a spreading machine according to claim 1, characterized in that: The woven fabric defect detection comprises the following steps: 321) Use frequency domain analysis to detect periodic defects; 322) Establish a normal fabric template and use template matching technology to detect and locate defective areas that deviate from the template; 323) In view of the color difference and uneven thickness of fabrics, the color and brightness anomalies are identified through color space transformation and brightness histogram analysis.

5. The method for online detection of fabric defects on a spreading machine according to claim 3, characterized in that: The stretch fabric defect detection comprises the following steps: 331) Use dynamic detection technology to detect defects in stretch fabrics and evaluate the changes in defects in stretch fabrics under stretching; 332) Use multi-scale analysis methods to detect defects at different scales.

6. The method for online defect detection of fabrics on a spreading machine according to claim 1, characterized in that: In the step 3), when a defect is detected, a contour extraction algorithm is used to perform statistical analysis on the size and shape of the defect in the surface image information, and each pixel point (x, y) in the defect area is traversed to obtain the upper left corner coordinates (min_x, min_y) and lower right corner coordinates (max_x, max_y) of the defect boundary box; and then the center coordinates (u, v) of the defect are calculated; and then the pixel coordinates of the defect are converted into the actual physical coordinates (X, Y) on the fabric surface according to the homography matrix M of the image acquisition device. w ,Y w ).

7. The method for online defect detection of fabrics on a spreading machine according to claim 1, characterized in that: The stored category and location information of the defect is transferred and used for neural network learning to improve the robustness of the defect detection algorithm.

Images