Fabric dyeing quality evaluation method based on image processing

Abstract

This invention relates to the field of fabric dyeing quality assessment technology. Specifically, it relates to a fabric dyeing quality assessment method based on image processing. The method includes the following steps: S1, acquiring standard dyeing images and fabric parameters, setting up image acquisition positions along the fabric transmission path, and dividing the acquisition path into acquisition segments based on adjacent image acquisition positions; S2, acquiring fabric image data from each image acquisition position and fabric video data from each acquisition segment. This invention constructs a correlation model between illumination parameter changes and color difference changes, calculates the initial illumination interference factor using a ratio method, and dynamically corrects it by combining the fabric reflectivity and texture structure parameters. Simultaneously, it separately calculates the background color difference shift data caused by fabric reflected light, achieving precise two-dimensional correction of illumination interference and background reflection interference. This effectively eliminates pseudo-anomalies caused by non-dyeing factors, truly restores the dyeing state of the fabric, and solves the problem of incomplete traditional single-interference correction.

Description

Technical Field

[0001] This invention relates to the field of fabric dyeing quality assessment technology, and more specifically, to a fabric dyeing quality assessment method based on image processing. Background Technology

[0002] Fabric dyeing quality assessment is a core part of the production process in the textile printing and dyeing industry. Its core purpose is to detect the color uniformity, color difference and dyeing defects of the fabric after dyeing, so as to ensure that the quality of the fabric leaving the factory meets the design standards and customer requirements.

[0003] Existing online inspection technologies suffer from motion blur due to the continuous movement of fabrics on production lines. Traditional high-speed shooting easily produces motion blur, leading to pixel-level registration errors and affecting the accuracy of color difference calculation. While some technologies attempt to increase the shooting frame rate, they do not fundamentally solve the motion interference problem and ignore the impact of fabric posture changes during movement on color acquisition. Furthermore, the lighting environment in industrial workshops is complex, with uneven spatial light distribution and temporal fluctuations in light intensity and color temperature. Simultaneously, the fabric's own reflected light can interfere with the conveyor belt background. Existing technologies mostly use fixed lighting compensation algorithms or single light source calibration, without dynamically correcting lighting interference based on the fabric's material, texture, and other parameters. This results in inaccurate calculation of the lighting interference factor and the presence of false anomalies in the color difference data. Therefore, an image processing-based fabric dyeing quality assessment method is proposed. Summary of the Invention

[0004] The purpose of this invention is to provide a fabric dyeing quality assessment method based on image processing to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, an image processing-based fabric dyeing quality assessment method is provided, comprising the following steps:

[0006] S1. Obtain the standard dyeing image and fabric parameters of the fabric, set up image acquisition positions on the fabric transmission path, and divide the acquisition segment according to the adjacent image acquisition positions.

[0007] S2. Collect fabric image data and fabric video data from each image acquisition location and each acquisition segment. Combine the standard dyeing image to perform color difference analysis on the image data from each acquisition location to obtain the fabric color difference data corresponding to each acquisition location.

[0008] S3. Locate abnormal areas on the fabric based on the fabric color difference data from all collection locations, track the abnormal areas using video data to lock their spatiotemporal trajectory, calculate the fabric color difference data of the abnormal areas in each collection segment by combining standard dyeing images, and calculate the color difference change data based on the color difference data of adjacent collection segment points.

[0009] S4. Obtain the illumination parameters at each position along the transmission path, combine the color difference change data to set illumination interference factors for different positions along the transmission path, then correct the illumination interference factors using fabric parameters, and simultaneously calculate the color difference offset data of the fabric before and after entering the scene at each image acquisition position.

[0010] S5. Based on the corrected illumination interference factor and color difference offset data, perform true color difference correction on the fabric color difference data of the collected road section and collection location, compare the corrected color difference data with the standard dyeing image, and determine whether the abnormal area is a true dyeing abnormality.

[0011] As a further improvement to this technical solution, in step S1, a communication connection is established with the fabric evaluation end to extract the standard dyeing image and fabric parameters of the fabric.

[0012] The standard dyeing images are the set of images that indicate the fabric has passed evaluation;

[0013] Meanwhile, a camera device installed on the fabric transport path is connected to the fabric evaluation end. The image acquisition position is set according to the number of cameras and the length of the fabric transport path. When the fabric moves to the image acquisition position, it pauses briefly so that the camera device can capture complete fabric image data before moving to the next acquisition position.

[0014] The fabric transport path between adjacent collection locations is designated as the collection segment.

[0015] As a further improvement to this technical solution, the setting of the image acquisition position satisfies that the distance between two adjacent image acquisition positions is less than the fabric length captured by the camera in a single frame.

[0016] The length of the acquisition segment is matched with the camera's shooting frame rate and the conveyor belt's running speed to ensure that the video data can completely capture the movement of the fabric between adjacent acquisition positions.

[0017] As a further improvement to this technical solution, in step S2, image data of the fabric at each collection location and video data of each collection segment are collected by a camera device.

[0018] The image data from each acquisition location is combined with the standard dyeing image and converted to the same color space. The color information is compared pixel by pixel and the color difference value of each pixel is calculated. Then, the color difference values ​​of all pixels at the same acquisition location are statistically processed to obtain the fabric color difference data corresponding to that acquisition location.

[0019] As a further improvement to this technical solution, in step S3, the color difference threshold is set according to the fabric evaluation end, and the fabric color difference data of each collection position is compared with the color difference threshold.

[0020] If the fabric color difference data exceeds the color difference threshold, the pixel is determined to be abnormal.

[0021] Conversely, if the fabric color difference data does not exceed the color difference threshold, the pixel is considered normal.

[0022] Then, the abnormal pixels are identified and noise is eliminated, and the set of connected pixels is defined as the abnormal region.

[0023] The target tracking is performed using a feature matching method. The color and texture features of the abnormal region are extracted, and the same abnormal region is matched in each frame of the video data and its spatiotemporal coordinates are recorded to form a spatiotemporal trajectory.

[0024] As a further improvement to this technical solution, based on the color and texture features, corresponding points on adjacent collection sections are selected, and the color difference data of the fabric in the abnormal area of ​​the two points is retrieved. The absolute value of the difference between the two is calculated, and this absolute value is the color difference change data between adjacent collection sections.

[0025] As a further improvement to this technical solution, in step S4, the light sensor installed on the fabric transport path is extracted through the fabric evaluation end, and then the light parameters are obtained based on the light sensor.

[0026] Illumination parameters include illuminance and color temperature parameters at various locations along the transmission path. These parameters are acquired by deploying illumination sensors at each acquisition location and along the acquisition path. The sensor sampling frequency is consistent with the camera acquisition frequency. The location of the illumination parameters along the transmission path is determined based on the coverage of the illumination sensors.

[0027] In the same location along the transmission path, a correlation is established between the change in illumination parameters and the change in color difference. The change in color difference at adjacent locations is used as the dependent variable, and the illumination parameters are used as the independent variable. The illumination interference factor is calculated using the ratio method, and the calculated illumination interference factor is then used for location setting.

[0028] Among them, the larger the ratio of the change in color difference to the change in illumination parameters, the higher the degree of interference of illumination on color difference at that location, and the larger the value of the illumination interference factor.

[0029] The reflectivity and texture parameters in the fabric parameters are retrieved. The reflectivity correction weight is set based on the fabric reflectivity, and the texture mask coefficient is generated based on the fabric texture structure. The reflectivity correction weight and texture mask coefficient are substituted into the set illumination interference factor to obtain the final corrected illumination interference factor.

[0030] As a further improvement to this technical solution, in step S4, background image data of the collection position before the fabric enters and background image data of the collection position after the fabric enters are collected respectively. Pixels of the corresponding background areas in the two types of images are extracted, the average color information of each group of pixels is calculated, and then the difference between the two groups of average color information is calculated by the color difference formula. This difference is the color difference offset data of the corresponding collection position.

[0031] As a further improvement to this technical solution, in step S5, based on the corrected light interference factor and color difference offset data, the fabric color difference data of the collection section and collection location are subjected to deviation elimination processing for light interference and background reflection interference to obtain corrected color difference data that reflects the true dyeing state of the fabric.

[0032] The corrected color difference data was compared with the standard dyed image;

[0033] If the corrected color difference data is the same as the standard dyeing image, the abnormal area is determined to have no color difference, and the abnormal area mark is removed.

[0034] Conversely, if the corrected color difference data is different from the standard dyeing image, it is determined that there is a color difference, and the abnormal area is confirmed as a true dyeing abnormality.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0036] 1. This image processing-based fabric dyeing quality assessment method constructs a correlation model between illumination parameter changes and color difference changes, calculates the initial illumination interference factor using the ratio method, and dynamically corrects it by combining the fabric reflectance coefficient and texture structure parameters. At the same time, it separately calculates the background color difference offset data caused by fabric reflected light, achieving precise correction of illumination interference and background reflection interference in two dimensions. This effectively removes pseudo-anomalies caused by non-dyeing factors, truly restores the dyeing state of the fabric, and solves the problem of incomplete correction of traditional single interference.

[0037] 2. In this image processing-based fabric dyeing quality assessment method, a combined strategy of color difference threshold comparison, morphological denoising, and dual-feature KCF tracking is adopted. First, isolated noise points are eliminated through erosion and dilation operations to ensure accurate positioning of abnormal areas. Then, based on the dual feature combination of color and texture features, combined with the KCF algorithm, stable tracking of abnormal areas is achieved, ensuring the continuity and accuracy of spatiotemporal trajectory. This makes the calculation of cross-segment color difference change data more valuable and provides accurate quantitative basis for the correlation of illumination interference. Attached Figure Description

[0038] Figure 1 This is a flowchart illustrating the fabric dyeing quality assessment method based on image processing according to the present invention.

[0039] Figure 2 This is a flowchart of S1 of the present invention;

[0040] Figure 3 This is a flowchart of S2 of the present invention;

[0041] Figure 4 This is a flowchart of S3 of the present invention;

[0042] Figure 5 This is a flowchart of S4 of the present invention;

[0043] Figure 6 This is a flowchart of S5 of the present invention. Detailed Implementation

[0044] 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.

[0045] Please see Figures 1-6 As shown, the purpose of this embodiment is to provide a fabric dyeing quality assessment method based on image processing, including the following steps:

[0046] S1. Obtain standard dyeing images and fabric parameters of the fabric, set up image acquisition positions on the fabric transmission path, and divide the acquisition path according to adjacent image acquisition positions; complete the basic data input and spatial layout of the acquisition system to provide hardware and data prerequisites for subsequent image / video acquisition and color difference analysis;

[0047] In S1, a communication connection is established with the fabric evaluation end to extract the standard dyeing image and fabric parameters of the fabric.

[0048] A two-way communication connection is established with the fabric evaluation end. Through this connection, two types of basic data are extracted from the fabric evaluation end database, including the standard dyeing image set of the fabric (composed of non-interference images of multiple sets of patterned fabrics) and fabric parameters (including key attributes such as material, reflectivity, and texture structure).

[0049] The standard dyeing images are the set of images that indicate the fabric has passed evaluation;

[0050] Meanwhile, a camera device installed on the fabric transport path is connected to the fabric evaluation end. The image acquisition position is set according to the number of cameras and the length of the fabric transport path. When the fabric moves to the image acquisition position, it pauses briefly so that the camera device can capture complete fabric image data before moving to the next acquisition position.

[0051] The fabric transport path between adjacent collection locations is designated as the collection segment.

[0052] Establish a communication connection between the camera device and the fabric evaluation end, and mark several image acquisition positions on the transmission path based on the number of camera devices deployed and the total length of the fabric transmission path.

[0053] Configure a linkage control program for the fabric transport system so that when the fabric moves to any image acquisition position, the transport system is briefly stopped.

[0054] Once the fabric has come to a complete stop, start the camera to capture images of the fabric at that location without motion blur.

[0055] After the shooting is completed, the transmission system resumes operation and transports the fabric to the next image acquisition location.

[0056] The image acquisition position is set such that the distance between two adjacent image acquisition positions is less than the length of fabric captured by the camera in a single frame;

[0057] The length of the data collection section is matched with the camera's frame rate and the conveyor belt's operating speed to ensure that the video data can completely capture the movement of the fabric between adjacent collection locations.

[0058] S2. Collect fabric image data and fabric video data from each image acquisition location and each acquisition segment. Combine the standard dyeing image to perform color difference analysis on the image data from each acquisition location to obtain the fabric color difference data corresponding to each acquisition location. Collect image / video dual-modal data and calculate the initial fabric color difference data to provide core basis for anomaly area location.

[0059] In S2, image data of the fabric at each collection location and video data of each collection section are collected using a camera device;

[0060] Control the acquisition frequency of the camera device to ensure that the acquisition time of image data corresponds one-to-one with the frame time of video data, and ensure the spatiotemporal synchronization of the two types of data;

[0061] The image data from each acquisition location is combined with a standard dyeing image and converted to the same color space. The color information is compared pixel by pixel, and the color difference value of each pixel is calculated. Then, the color difference values ​​of all pixels at the same acquisition location are statistically processed to obtain the fabric color difference data corresponding to that acquisition location. The steps are as follows:

[0062] Retrieve the standard dyeing image set, and simultaneously convert the static fabric image data and standard dyeing image data from each acquisition location to the CIELab color space (the best standard space for industrial color difference analysis). During the conversion process, eliminate the influence of equipment color gamut differences and illumination color gamut distortion to ensure that the color representation system of the acquired images and the standard images is consistent. Then, perform pixel-level registration between the converted acquired images and the standard images to ensure that the same pixel in the two images corresponds to the same physical location on the fabric. Extract the Lab color components of the corresponding pixels after registration, calculate the color difference value pixel by pixel, and record the color difference result of each pixel.

[0063] For the color difference values ​​of all pixels at the same sampling location, statistical features are extracted. The mean (reflecting the overall color difference level of the fabric at that location) and the extreme value (reflecting the maximum color difference deviation of the fabric at that location) are selected as the fabric color difference data corresponding to that sampling location and uploaded synchronously to the fabric evaluation terminal. The formula is as follows:

[0064] ;

[0065] in, This represents the color difference value for a single pixel. A larger value indicates a greater color deviation between that pixel and the standard image. To collect the difference in luminance component between the luminance value of a pixel in the image and the luminance value of the corresponding pixel in the standard image, This represents the difference between the red and green components; a positive value indicates a red tint, and a negative value indicates a green tint. This represents the difference between the yellow and blue components; a positive value indicates a yellowish tint, and a negative value indicates a bluish tint.

[0066] ;

[0067] in, Let N be the average color difference at a certain sampling location, and N be the total number of pixels in the image at that sampling location. Let be the color difference value of the i-th pixel;

[0068] ;

[0069] in, It represents the maximum value of the color difference among all pixels at a certain sampling location.

[0070] S3. Based on the fabric color difference data from all collection locations, locate abnormal areas on the fabric. Use video data to track the abnormal areas to lock their spatiotemporal trajectory. Combine standard dyeing images to calculate the fabric color difference data of the abnormal areas in each collection segment. Calculate the color difference change data based on the color difference data of adjacent collection segment points. Identify abnormal areas from the initial color difference data, lock the spatiotemporal trajectory of the abnormal areas through video tracking, and calculate the color difference change data across segments, providing key support for subsequent correlation of light interference.

[0071] In S3, the color difference threshold is set according to the fabric evaluation end (the threshold is dynamically configured according to the fabric type and quality grade, and the threshold is more stringent for high value-added fabrics), and the fabric color difference data from each collection location is compared with the color difference threshold.

[0072] If the fabric color difference data exceeds the color difference threshold, the pixel is determined to be abnormal.

[0073] Conversely, if the fabric color difference data does not exceed the color difference threshold, the pixel is considered normal.

[0074] Then, the abnormal pixels are identified and noise is eliminated, and the set of connected pixels is defined as the abnormal region.

[0075] First, perform erosion operation to eliminate isolated noise points (such as false anomalies caused by dust or sensor noise), then perform dilation operation to restore the boundary of the real anomaly region, delineate the set of interconnected anomaly pixels after processing as independent anomaly regions, and record the initial position coordinates and area of ​​each anomaly region.

[0076] The target tracking is performed using a feature matching method. The color and texture features of the abnormal region are extracted, and the same abnormal region is matched in each frame of the video data and its spatiotemporal coordinates are recorded to form a spatiotemporal trajectory.

[0077] The color features (Lab mean) and texture features (gray-level co-occurrence matrix entropy value) of each abnormal region are extracted as tracking features. At the same time, the KCF target tracking algorithm is used to match the same abnormal region in each frame of the video data of the acquired road segment, record the coordinate information of the abnormal region at different times and locations, and generate the spatiotemporal trajectory of the abnormal region.

[0078] Based on color and texture characteristics, corresponding points on adjacent collection sections are selected, and the color difference data of fabrics in abnormal areas of the two points are retrieved. The absolute value of the difference between the two is calculated, and this absolute value is the color difference change data between adjacent collection sections.

[0079] Based on the spatiotemporal trajectory, corresponding points (fabric areas at the same physical location) of abnormal areas are selected in adjacent collection segments. Color difference data of abnormal areas at two points are retrieved, and the absolute value of the difference is calculated to obtain color difference change data between adjacent collection segments.

[0080] S4. Obtain the illumination parameters at each location along the transmission path, combine the color difference change data to set illumination interference factors for different locations along the transmission path, and then correct the illumination interference factors using fabric parameters. At the same time, calculate the color difference offset data of the fabric before and after entering the scene at each image acquisition location; quantify the two core interference factors of illumination interference and background reflection interference, and generate key parameters for real color difference correction (corrected illumination interference factor + color difference offset data).

[0081] In S4, the illumination sensor installed on the fabric transport path is extracted through the fabric evaluation end, and then the illumination parameters are obtained based on the illumination sensor.

[0082] A communication connection with the light sensor is established through the fabric evaluation end. Illuminance-color temperature integrated sensors are deployed at each image acquisition location and at the beginning, end and middle of each acquisition section. The sensor sampling frequency is set to be consistent with the acquisition frequency of the camera device. Illuminance parameters and color temperature parameters at each location are collected in real time. Based on the sensor deployment location, a one-to-one correspondence between the light parameters and the transmission path location is established.

[0083] Illumination parameters include illuminance and color temperature parameters at various locations along the transmission path. These parameters are acquired by deploying illumination sensors at each acquisition location and along the acquisition path. The sensor sampling frequency is consistent with the camera acquisition frequency. The location of the illumination parameters along the transmission path is determined based on the coverage of the illumination sensors.

[0084] In the same location along the transmission path, a correlation is established between the change in illumination parameters and the change in color difference. The change in color difference at adjacent locations is used as the dependent variable, and the illumination parameters are used as the independent variable. The illumination interference factor is calculated using the ratio method, and the calculated illumination interference factor is then used for location setting.

[0085] For adjacent locations along the transmission path (including adjacent acquisition locations and adjacent sampling points within the acquisition segment), the changes in illuminance and color temperature are calculated respectively. Based on the calculation results, a dataset of changes in illumination parameters at each location is formed, providing a quantitative basis for modeling illumination interference factors.

[0086] Then, a correlation model between the change in illumination parameters and the change in color difference data is established. The change in color difference data at adjacent positions is used as the dependent variable, and the change in illumination parameters is used as the independent variable. The initial illumination interference factor value is calculated to be positively correlated with the degree of interference of illumination on color difference by the ratio method. At the same time, the initial factor is bound to the corresponding position to complete the factor setting at each position of the transmission path.

[0087] The larger the ratio of the change in color difference to the change in illumination parameters, the greater the interference of illumination on the color difference at that location, and the larger the value of the illumination interference factor, as shown in the formula below:

[0088] ;

[0089] in, The initial light interference factor. This represents the color difference variation data between adjacent positions. The change in illuminance This is the color temperature influence weighting coefficient (with a value of 0.01 to 0.05, used to balance the influence of illuminance and color temperature on color difference). This represents the change in color temperature.

[0090] Retrieve the reflectivity and texture parameters from the fabric parameters. Set a reflectivity correction weight based on the reflectivity (high reflectivity fabrics have a higher weight than matte fabrics). Generate a texture mask coefficient based on the fabric texture structure (the more complex the texture, the larger the texture mask coefficient value). Substitute the reflectivity correction weight and texture mask coefficient into the set illumination interference factor to obtain the corrected final illumination interference factor, as shown in the following formula:

[0091] ;

[0092] in, This is the corrected final illumination interference factor. The reflectivity of the fabric (value range 0-1, e.g., pure cotton fabric = 0.2, polyester fabric = 0.6). This is the texture mask coefficient (1 for areas without texture, and 1.2-1.5 for complex texture areas such as knitted / denim).

[0093] In S4, background image data of the location before the fabric enters and background image data of the location after the fabric enters are collected respectively. Pixels of the corresponding background areas in the two types of images are extracted, the average color information of each group of pixels is calculated, and the difference between the two groups of average color information is calculated using the color difference formula. This difference is the color difference offset data of the corresponding collection location.

[0094] S5. Based on the corrected illumination interference factor and color difference offset data, perform true color difference correction on the fabric color difference data of the collected road sections and locations. Compare the corrected color difference data with the standard dyeing image to determine whether the abnormal area is a true dyeing abnormality. Use the interference parameters of S4 to correct the initial color difference data to obtain the true dyeing color difference of the fabric. Finally, determine the attribute of the abnormal area (false illumination abnormality / true dyeing abnormality) to complete the quality assessment.

[0095] In S5, based on the corrected light interference factor and color difference offset data, the fabric color difference data of the collected road section and collection location are processed to eliminate the deviation of light interference and background reflection interference, so as to obtain the corrected color difference data that reflects the true dyeing state of the fabric.

[0096] The model constructs a dual-interference correction model for illumination and background reflection. The original color difference data is substituted into the model to simultaneously eliminate background shift interference caused by illumination fluctuations and fabric reflection. The corrected color difference data, reflecting the true dyeing state of the fabric, is calculated using the following formula:

[0097] ;

[0098] in, This is the corrected, true color difference data. To collect raw fabric color difference data for road sections / locations, This represents the actual illuminance value at the current location. The reference illuminance value for a standard light source. This is for color difference offset data at the acquisition location.

[0099] The corrected color difference data was compared with the standard dyed image;

[0100] If the corrected color difference data is the same as the standard dyeing image, the abnormal area is determined to have no color difference, and the abnormal area mark is removed.

[0101] Conversely, if the corrected color difference data is different from the standard dyeing image, it is determined that there is a color difference, and the abnormal area is confirmed as a real dyeing abnormality.

[0102] At the same time, based on the required characteristics of the material, a set deviation threshold can be used as the judgment label;

[0103] If the deviation between the corrected color difference data and the standard staining image is greater than the deviation threshold, then there is a color difference, and the abnormal area is confirmed as a true staining abnormality. Conversely, if the deviation is less than the deviation threshold, then there is no color difference.

[0104] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A fabric dyeing quality assessment method based on image processing, characterized in that: Includes the following steps: S1. Obtain the standard dyeing image and fabric parameters of the fabric, set up image acquisition positions on the fabric transmission path, and divide the acquisition segment according to the adjacent image acquisition positions. S2. Collect fabric image data and fabric video data from each image acquisition location and each acquisition segment. Combine the standard dyeing image to perform color difference analysis on the image data from each acquisition location to obtain the fabric color difference data corresponding to each acquisition location. S3. Locate abnormal areas on the fabric based on the fabric color difference data from all collection locations, track the abnormal areas using video data to lock their spatiotemporal trajectory, calculate the fabric color difference data of the abnormal areas in each collection segment by combining standard dyeing images, and calculate the color difference change data based on the color difference data of adjacent collection segment points. S4. Obtain the illumination parameters at each position along the transmission path, combine the color difference change data to set illumination interference factors for different positions along the transmission path, then correct the illumination interference factors using fabric parameters, and simultaneously calculate the color difference offset data of the fabric before and after entering the scene at each image acquisition position. S5. Based on the corrected illumination interference factor and color difference offset data, perform true color difference correction on the fabric color difference data of the collected road section and collection location, compare the corrected color difference data with the standard dyeing image, and determine whether the abnormal area is a true dyeing abnormality.

2. The fabric dyeing quality assessment method based on image processing according to claim 1, characterized in that: In step S1, a communication connection is established with the fabric evaluation end to extract the standard dyeing image and fabric parameters of the fabric. The standard dyeing images are the set of images that indicate the fabric has passed evaluation; Meanwhile, a camera device installed on the fabric transport path is connected to the fabric evaluation end. The image acquisition position is set according to the number of cameras and the length of the fabric transport path. When the fabric moves to the image acquisition position, it pauses briefly so that the camera device can capture complete fabric image data before moving to the next acquisition position. The fabric transport path between adjacent collection locations is designated as the collection segment.

3. The fabric dyeing quality assessment method based on image processing according to claim 2, characterized in that: The image acquisition position is set such that the distance between two adjacent image acquisition positions is less than the fabric length captured by the camera in a single frame. The length of the acquisition segment is matched with the camera's shooting frame rate and the conveyor belt's running speed to ensure that the video data can completely capture the movement of the fabric between adjacent acquisition positions.

4. The fabric dyeing quality assessment method based on image processing according to claim 1, characterized in that: In step S2, image data of the fabric at each collection location and video data of each collection segment are collected using a camera device. The image data from each acquisition location is combined with the standard dyeing image and converted to the same color space. The color information is compared pixel by pixel and the color difference value of each pixel is calculated. Then, the color difference values ​​of all pixels at the same acquisition location are statistically processed to obtain the fabric color difference data corresponding to that acquisition location.

5. The fabric dyeing quality assessment method based on image processing according to claim 4, characterized in that: In step S3, the color difference threshold is set according to the fabric evaluation end, and the fabric color difference data of each collection location is compared with the color difference threshold. If the fabric color difference data exceeds the color difference threshold, the pixel is determined to be abnormal. Conversely, if the fabric color difference data does not exceed the color difference threshold, the pixel is considered normal. Then, the abnormal pixels are identified and noise is eliminated, and the set of connected pixels is defined as the abnormal region. The target tracking is performed using a feature matching method. The color and texture features of the abnormal region are extracted, and the same abnormal region is matched in each frame of the video data and its spatiotemporal coordinates are recorded to form a spatiotemporal trajectory.

6. The fabric dyeing quality assessment method based on image processing according to claim 5, characterized in that: Based on the color and texture features, corresponding points on adjacent collection sections are selected, and the color difference data of the fabric in the abnormal area of ​​the two points is retrieved. The absolute value of the difference between the two is calculated, and this absolute value is the color difference change data between adjacent collection sections.

7. The fabric dyeing quality assessment method based on image processing according to claim 1, characterized in that: In step S4, the illumination sensor installed on the fabric transport path is extracted through the fabric evaluation end, and then the illumination parameters are obtained based on the illumination sensor. Illumination parameters include illuminance and color temperature parameters at various locations along the transmission path. These parameters are acquired by deploying illumination sensors at each acquisition location and along the acquisition path. The sensor sampling frequency is consistent with the camera acquisition frequency. The location of the illumination parameters along the transmission path is determined based on the coverage of the illumination sensors. In the same location along the transmission path, a correlation is established between the change in illumination parameters and the change in color difference. The change in color difference at adjacent locations is used as the dependent variable, and the illumination parameters are used as the independent variable. The illumination interference factor is calculated using the ratio method, and the calculated illumination interference factor is then used for location setting. Among them, the larger the ratio of the change in color difference to the change in illumination parameters, the higher the degree of interference of illumination on color difference at that location, and the larger the value of the illumination interference factor. The reflectivity and texture parameters in the fabric parameters are retrieved. The reflectivity correction weight is set based on the fabric reflectivity, and the texture mask coefficient is generated based on the fabric texture structure. The reflectivity correction weight and texture mask coefficient are substituted into the set illumination interference factor to obtain the final corrected illumination interference factor.

8. The fabric dyeing quality assessment method based on image processing according to claim 1, characterized in that: In step S4, background image data of the collection position before the fabric enters and background image data of the collection position after the fabric enters are collected respectively. Pixels of the corresponding background areas in the two types of images are extracted, the average color information of each group of pixels is calculated, and the difference between the two groups of average color information is calculated by the color difference formula. This difference is the color difference offset data of the corresponding collection position.

9. The fabric dyeing quality assessment method based on image processing according to claim 1, characterized in that: In step S5, based on the corrected light interference factor and color difference offset data, the fabric color difference data of the collected road section and collection location are subjected to deviation elimination processing for light interference and background reflection interference to obtain corrected color difference data that reflects the true dyeing state of the fabric. The corrected color difference data was compared with the standard dyed image; If the corrected color difference data is the same as the standard dyeing image, the abnormal area is determined to have no color difference, and the abnormal area mark is removed. Conversely, if the corrected color difference data is different from the standard dyeing image, it is determined that there is a color difference, and the abnormal area is confirmed as a true dyeing abnormality.

Images