A yarn uniformity analysis method based on image processing

Abstract

The application discloses a yarn uniformity analysis method based on image processing and relates to the technical field of image data processing, which comprises the following steps: through a preliminary feature distribution map, a convolutional neural network model is used to perform deep learning analysis on yarn surface features, adaptive feature weighting processing is performed on complex texture areas, and a more accurate third feature distribution map is obtained; if the feature values of some areas in the third feature distribution map exceed a preset abnormal threshold value, local magnification analysis is performed on the areas, a fine defect boundary is further extracted in combination with an edge detection algorithm, and a defect candidate area set is obtained; according to multidimensional evaluation data, a visual analysis report is generated, a heat map technology is applied to intuitively present yarn uniformity distribution, specific position information is marked for defect high-incidence areas, and a final detection analysis result is output. The application can effectively solve the problems of yarn image acquisition and defect detection under complex illumination, and improve the accuracy and efficiency of yarn uniformity detection.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of image data processing, and particularly relates to a yarn uniformity analysis method based on image processing. BACKGROUND

[0002] In the field of textile industry, the uniformity of yarn is a core element that determines the quality and performance of fabric, directly related to the market competitiveness and production efficiency of products. Research and innovation in this field are of great significance to promote the automation and intelligentization of the industry. However, the current mainstream detection methods have significant limitations. Traditional manual visual inspection is not only inefficient, but also easily affected by the subjective experience of operators, making it difficult to meet the needs of large-scale production. The detection methods based on physical devices often need to directly contact the yarn, which has the risk of damaging the sample, and the comprehensive evaluation ability of multiple quality parameters is insufficient, making it difficult to meet the requirements of complex production scenarios.

[0003] Under this background, yarn uniformity detection faces many technical challenges. The first and foremost is how to accurately obtain the surface feature information of the yarn in a complex environment. Due to uneven lighting conditions or shadow interference in the production site, traditional image acquisition methods often cannot clearly present the true form of the yarn, leading to increased errors in subsequent analysis. This instability of image quality further affects the reliability of feature extraction, especially when facing yarns with complex textures or color differences, conventional methods are difficult to accurately distinguish normal structural features from defect features, thus leading to deviation or even failure of the detection results. This series of problems makes it a difficult problem to achieve efficient and comprehensive uniformity evaluation. SUMMARY

[0004] To solve the above technical problems, the application provides a yarn uniformity analysis method based on image processing, which can effectively solve the problems of yarn image acquisition and defect detection under complex lighting, and improve the accuracy and efficiency of yarn uniformity detection.

[0005] To achieve the above purpose, the application provides a yarn uniformity analysis method based on image processing, comprising:

[0006] Obtaining initial yarn image data, and obtaining a first yarn image according to the initial yarn image data;

[0007] According to the brightness distribution characteristics of the first yarn image, performing local contrast enhancement processing on the uneven lighting area, and performing gray scale normalization on the shadow interference area of the enhanced image to obtain a second yarn image with balanced brightness;

[0008] According to the texture feature difference of the second yarn image, the image is layered into multiple frequency subbands through image decomposition technology of wavelet transform, the texture detail information in the high frequency subband is extracted, and a preliminary feature distribution map of the yarn surface is obtained;

[0009] Through the preliminary feature distribution map, the yarn surface features are analyzed in depth, the adaptive feature weighting processing is performed on the analyzed complex texture region, and a more accurate third feature distribution map is obtained.

[0010] The region with a feature value exceeding a preset abnormal threshold in the third feature distribution map is analyzed locally, and the fine defect boundary is further extracted by combining an edge detection algorithm, and a final defect candidate region set is obtained.

[0011] According to the defect candidate region set, the candidate regions are grouped and processed, the feature similarity index of each group of regions is calculated, if the similarity index is lower than a preset threshold, the non-defect region is removed, and the final defect positioning result is determined.

[0012] Through the final defect positioning result, the uniformity evaluation score is calculated by combining the overall feature distribution of the yarn, the defect density and feature deviation of different regions are comprehensively quantified, and multi-dimensional evaluation data of yarn uniformity is obtained.

[0013] According to the multi-dimensional evaluation data, a visual analysis report is generated, the yarn uniformity distribution is intuitively presented, the specific position information of the defect high incidence area is marked, and the final detection analysis result is obtained.

[0014] The technical effect of the present application is that the present application discloses a yarn uniformity analysis method based on image processing, which is aimed at the problem of yarn image acquisition under complex lighting conditions, and obtains an initial image through a multi-light source adaptive adjustment system and performs light balance processing. BRIEF DESCRIPTION OF DRAWINGS

[0015] The accompanying drawings, which form a part of this application, are intended to provide further understanding of the application and serve to explain the illustrative embodiments of the present application and their descriptions, and do not constitute improper limitations on the present application. In the drawings:

[0016] Figure 1 FIG. 1 is a flowchart of a yarn uniformity analysis method based on image processing according to an embodiment of the present application. DETAILED DESCRIPTION

[0017] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0018] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in a different order.

[0019] As shown in FIG. 1, the present embodiment provides a yarn uniformity analysis method based on image processing, which comprises the following steps. Figure 1

[0020] According to the brightness distribution characteristics of the first yarn image, a local contrast enhancement processing is performed on the uneven illumination area, and a gray scale normalization is performed on the shadow interference area of the enhanced image to obtain a second yarn image with balanced brightness.

[0021] According to the texture feature difference of the second yarn image, the image is layered into a plurality of frequency subbands through the image decomposition technology of wavelet transform, the texture detail information in the high frequency subband is extracted, and a preliminary feature distribution map of the yarn surface is obtained.

[0022] Through the preliminary feature distribution map, a deep learning analysis is performed on the yarn surface features, an adaptive feature weighting processing is performed on the analyzed texture complex area, and a more accurate third feature distribution map is obtained.

[0023] A local magnification analysis is performed on the area in the third feature distribution map whose feature value exceeds a preset abnormal threshold, and a fine defect boundary is further extracted in combination with an edge detection algorithm to obtain a final defect candidate area set.

[0024] According to the defect candidate area set, a grouping processing is performed on the candidate areas, a feature similarity index is calculated for each group of areas, and if the similarity index is lower than a preset distinguishing threshold, a non-defect area is removed, and a final defect positioning result is determined.

[0025] Through the final defect positioning result, in combination with the overall feature distribution of the yarn, a uniformity evaluation score is calculated, a comprehensive quantification is performed on the defect density and feature deviation of different areas, and a multi-dimensional evaluation data of the yarn uniformity is obtained.

[0026] ​According to the multi-dimensional evaluation data, a visual analysis report is generated to intuitively present the yarn uniformity distribution, specific position information is marked for a defect high-occurrence area, and a final detection analysis result is obtained.

[0027] Further, the initial yarn image data acquisition includes:

[0028] According to the multi-light source adaptive adjustment mechanism, for the yarn features under a complex light scene, a multi-angle illumination mode is adopted, environment interference data is acquired through a preset light intensity sensor, the light source output intensity is adjusted in real time by combining a light intensity dynamic module, and the initial yarn image data is acquired.

[0029] Further, the first yarn image obtained according to the initial yarn image data includes:

[0030] Through the initial yarn image data, an initial image is acquired under multi-angle illumination conditions by using an image acquisition device, for the environmental interference factors, if it is detected that the light distribution is uneven by more than a preset threshold, the light source output is further adjusted by the light intensity dynamic module, and the first yarn image is obtained.

[0031] Specifically, under a complex light scene, in order to accurately capture the yarn features, the multi-light source adaptive adjustment mechanism is particularly important. For the multi-angle illumination mode, a plurality of light sources can be arranged, such as setting an adjustable light intensity LED lamp at the top, side and bottom of the yarn detection area, respectively illuminating the yarn surface at an angle of 45 degrees, 90 degrees and 135 degrees. This design can cover different sides of the yarn and reduce shadow interference. Combined with the light intensity sensor, assuming that the ambient light intensity is 200 lux and the preset optimal detection light intensity is 500 lux, the system will detect the environmental interference data in real time, and adjust the light source output to a supplementary light intensity of 300 lux through the light intensity dynamic module, thereby forming a preliminary balanced light distribution data. This way effectively avoids overexposure or underexposure, improving the stability of image acquisition. Specifically, based on the balanced light distribution data, an image acquisition device such as a high-resolution industrial camera can acquire an initial image under multi-angle illumination conditions. If it is detected that the light distribution is uneven, for example, the top light intensity is 550 lux and the side light intensity is only 450 lux, which exceeds the preset threshold of 50 lux deviation, the system will further adjust through the light intensity dynamic module to increase the side light source output to 500 lux, forming an optimized first yarn image. This dynamic adjustment mechanism can significantly improve the uniformity of the image, laying a foundation for subsequent feature extraction.

[0032] Further, the second yarn image with balanced brightness includes:

[0033] According to the brightness distribution characteristics of the first yarn image, a local contrast enhancement processing is performed on the uneven illumination area, and a preliminary adjusted intermediate yarn image is obtained by adjusting the pixel value distribution range;

[0034] For the shadow interference area in the intermediate yarn image, the gray value of the local area is standardized. If the gray value of the shadow interference area deviates from the preset threshold value, the shadow interference area is adjusted again to obtain a transition yarn image with balanced gray scale;

[0035] By detecting the overall brightness distribution of the transition yarn image, the area with uneven illumination is refined to obtain an optimized yarn image with more uniform brightness distribution;

[0036] According to the brightness characteristics of the optimized yarn image, the gray value and contrast of the optimized yarn image are finally detected. If it is found that the local area still does not meet the preset threshold value, the local adjustment tool is used for fine tuning to obtain a second yarn image with balanced brightness.

[0037] Specifically, for the shadow interference area in the intermediate yarn image, the application of the gray scale normalization tool is particularly important. The shadow area often causes the gray value to be low, affecting subsequent analysis. Assuming that the average gray value of a certain shadow area is 60 and the preset threshold value is 80, the tool will standardize the gray value of this area and narrow the gap with other areas. If it still deviates from the threshold value after the first adjustment, for example, it is adjusted to 75, it will be adjusted again, and finally the gray value is close to 80, forming a transition yarn image with balanced gray scale. This method can effectively reduce the interference of shadows on the overall consistency of the image. In the overall brightness distribution detection of the transition yarn image, the fine tuning of the brightness balancing tool is crucial. Assuming that the detection finds that the edge area of the image has a brightness value of 120, while the center area has a brightness value of 180, there is a significant unevenness. The tool will adjust the brightness of the edge area, for example, to 150, while appropriately reducing the brightness of the center area to 160, to achieve the uniformity of the overall brightness distribution, and obtain an optimized yarn image. This fine tuning helps to ensure that the details of each area of the image are more consistent. For the brightness characteristic verification of the optimized yarn image, the image verification tool will finally detect the overall gray scale and contrast. Assuming that the preset gray value range is 100 to 200, and the gray value of a certain local area is only 90, the tool will mark this area as unqualified. Subsequently, the local adjustment tool will fine-tune the area, for example, by raising the pixel value to 105, to ensure that it meets the preset threshold value, and finally determine the second yarn image. This verification and fine-tuning mechanism can further ensure the image quality and provide a reliable foundation for subsequent yarn feature analysis.

[0038] Further, obtaining the preliminary feature distribution map of the yarn surface includes:

[0039] According to the texture characteristics of the second yarn image, the image is subjected to multi-level segmentation processing, different frequency components are separated, high-frequency sub-band data is obtained, and a frequency layered image after preliminary decomposition is obtained;

[0040] For the high-frequency sub-band in the frequency layered image, the texture details are subjected to separation processing, if the signal strength of the high-frequency sub-band is detected to exceed a preset threshold, enhancement processing is performed, and a set of detail features of the yarn surface is determined;

[0041] By distributing analysis on the set of detail features, the extracted texture details are associated and matched with the surface characteristics of the yarn, the distribution law corresponding to the surface characteristics is obtained, and a preliminary feature distribution map of the yarn surface is obtained.

[0042] Specifically, when processing the texture characteristics of the second yarn image, the image can be subjected to multi-level segmentation processing by an image decomposition tool. The core of this method is to decompose the image into different frequency components in order to separate the high-frequency sub-band data. Assuming that the yarn image contains surface fine texture and background noise, the image can be divided into high-frequency and low-frequency parts by the decomposition tool, the high-frequency part usually corresponds to the detail texture of the yarn surface, and the low-frequency part more reflects the overall background information. After decomposition, a frequency layered image is obtained, which lays a foundation for subsequent detail extraction. For the high-frequency sub-band part in the frequency layered image, the application of the detail extraction tool is particularly important. Assuming that the signal strength range of the high-frequency sub-band is preset to be 50, and the signal strength detected in a certain region reaches 70, it indicates that the region contains more obvious texture details. At this time, the tool will perform enhancement processing on this part to highlight the fine features of the yarn surface, such as the roughness or breaking point of the fiber, to form a set of detail features. This way can effectively capture the key information of the yarn surface and provide support for subsequent analysis. When distributing analysis is performed on the set of detail features, the feature mapping tool is used to associate and match the extracted texture details with the surface characteristics of the yarn. Assuming that it is found through analysis that the texture details of a certain region show periodic distribution, which may be caused by regular tension changes during the weaving process of the yarn. The tool will map this periodic feature to the surface characteristics to form a preliminary surface feature distribution map. This association and matching helps to reveal the potential law of the yarn surface and provides a basis for quality detection.

[0043] Further, obtaining a more accurate third feature distribution map includes:

[0044] According to the preliminary feature distribution map, a region division processing is performed on the texture complex region of the yarn surface, surface detail data of at least one target region is obtained, and a set of divided region details is obtained;

[0045] For the region detail set, the texture distribution of each region is processed separately. If the texture complexity of the region is detected to exceed a preset threshold, the region is adaptively adjusted to determine a weighted feature set;

[0046] The weighted feature set is used to process the feature matching on the surface of the yarn to obtain more accurate surface detail distribution data and obtain an optimized feature mapping result;

[0047] According to the optimized feature mapping result, the distribution map of the surface of the yarn is updated, the surface details of different regions are integrated, and a more accurate third feature distribution map is obtained.

[0048] Specifically, when processing the preliminary distribution map of the texture features of the surface of the yarn, the complex region can be divided by an image segmentation tool. Assuming that the texture density of some regions in the surface image of the yarn is high, showing interlaced or stacked characteristics, the tool will divide these regions into multiple independent small block regions according to the gray level change or boundary strength of the texture. For example, the gray value of a region changes in the range of 100 to 150, which is obviously higher than the average value 80 of the surrounding region, and is marked as a target region, and then the surface detail data thereof is obtained to form a region detail set. This division method helps to divide the complex texture region into smaller units for subsequent analysis. For the region detail set, the application of the feature extraction tool can further separate the texture distribution of each region. Assuming that the preset texture complexity threshold is 60, and the complexity detection value of a region is 75, which exceeds the threshold, the tool will start the adaptive adjustment mechanism, increase the analysis weight of the region, for example, adjust the weight from the default 1.0 to 1.5, highlight the texture features, and form a weighted feature set. This weighting process can make the subsequent analysis pay more attention to the details of the complex region, ensuring that important information is not ignored. When using the weighted feature set for in-depth analysis, the in-depth analysis tool will process the feature matching on the surface of the yarn. Assuming that the texture features of a region show a specific directional distribution, the tool will compare it with the known yarn surface feature database to confirm whether it is a common weaving texture feature, and then generate more accurate surface detail distribution data to form an optimized feature mapping result. This matching process can effectively improve the accuracy of the feature distribution and provide a reliable basis for subsequent integration.

[0049] Further, obtaining the final defect candidate region set includes:

[0050] According to the third feature distribution map, a target region exceeding a preset abnormal threshold is locally magnified to obtain magnified surface detail data and obtain a magnified detail view;

[0051] According to the amplified detail view, boundary recognition processing is performed on the fine defect, and if the clarity of the defect boundary is detected to be lower than a preset threshold, secondary sharpening processing is performed to determine the clear defect boundary range;

[0052] According to the clear defect boundary range, the defect region is divided to extract at least one independent candidate region, thereby obtaining a preliminary defect candidate region set;

[0053] According to the preliminary defect candidate region set, the surface details and distribution map of each candidate region are matched, and if the matching degree is lower than a preset threshold, detail extraction supplement is performed to obtain a final defect candidate region set.

[0054] Specifically, when processing the third feature distribution map of the yarn surface feature, for the region with abnormal feature value, the image zoom tool can be used to locally magnify the target region that exceeds the preset abnormal threshold. Assuming that the preset feature value abnormal threshold is 120, and the feature value of a certain region reaches 150, which obviously exceeds the threshold, the tool will magnify the image of this region to 2 times the original size, so that the surface detail data can be observed more clearly. This magnification processing can make the fine texture changes or potential defects more intuitive, and provide more rich visual information for subsequent analysis. For the magnified detail view, the application of the edge detection tool can help identify the boundary of the fine defect. Assuming that the defect boundary clarity threshold of a certain region is preset to 80%, and the actual detection value is only 60%, the tool will start secondary sharpening processing, which will adjust the image contrast or enhance the edge pixel value to improve the clarity to more than 85%, and determine the clear defect boundary range. This processing method can ensure the accuracy of the defect boundary and avoid misjudgment caused by fuzzy boundary.

[0055] Further, determining the final defect positioning result includes:

[0056] According to the final defect candidate region set, the candidate regions are preliminarily grouped, the surface feature data of the regions in each group is obtained, at least one feature vector is extracted therefrom, and a preliminary grouping result is obtained;

[0057] According to the preliminary grouping result, similarity calculation is performed on the feature vectors in the group, if the calculated similarity index is lower than a preset threshold, the corresponding region is marked as a non-defect region and removed, and a screened region set is determined;

[0058] Through the screened region set, the surface detail depth of the region is compared to obtain detail feature data, and it is judged whether it meets the target defect standard, thereby obtaining a candidate region set that meets the standard;

[0059] According to the set of candidate regions meeting the standard, boundary confirmation processing is performed on the regions, and if the boundary definition reaches a preset threshold value, the region is confirmed as a final defect position, and a final defect positioning result is obtained.

[0060] Specifically, in the textile industry, when processing surface defects of yarns, analysis of the set of defect candidate regions is particularly important. Using clustering tools to preliminarily group the candidate regions can classify regions with similar surface features into one category. Assuming that 10 candidate regions are identified in a yarn surface image, they are divided into 3 groups by the clustering tool according to texture density and color distribution. The regions in each group will further extract surface feature data such as texture roughness and color difference value to form a feature vector, laying a foundation for subsequent analysis. For the preliminary grouping result, the feature comparison tool will calculate the similarity of the feature vectors in each group. If the preset differentiation threshold is 70%, and the similarity of two regions in a group is only 50%, it indicates that they may not belong to the same defect type, and the tool will mark the region with low similarity as a non-defect region and exclude it. Assuming that there are 5 regions in a group, 2 of which are excluded, the remaining 3 regions form a screened region set. This approach helps to focus on regions with higher defect potential. On the basis of the screened region set, the image detail extraction tool will deeply compare the surface details of each region. Assuming that the texture breakage degree in the detail feature data of a region reaches 80% of the preset standard, it is judged to meet the target defect standard and is included in the set of candidate regions meeting the standard. Otherwise, if the detail feature of a region is only 40%, it will be excluded. This deep comparison can more accurately screen out potential defect regions.

[0061] Further, the multi-dimensional evaluation data of yarn uniformity includes:

[0062] According to the final defect positioning result, the yarn surface is divided into regions, at least one feature parameter is extracted for each region, and a region feature distribution set is obtained;

[0063] The defect density of each region is quantitatively processed through the region feature distribution set, and if the defect density is higher than a preset threshold value, it is marked as a high-density region to determine a high-density region set;

[0064] According to the high-density region set, a feature extraction tool is used to calculate the feature deviation of each region to obtain feature deviation data, and if the feature deviation exceeds a preset threshold value, it is marked as a deviation abnormal region to obtain a deviation abnormal region set;

[0065] Through the deviation abnormal region set and the region feature distribution set, a weighted average calculation is performed to comprehensively process the yarn features and defect density of the region, and multi-dimensional evaluation data of yarn uniformity is obtained.

[0066] Specifically, in the business scenario of yarn surface defect analysis, it is particularly important to divide the area and extract the feature parameters for the final defect positioning result. Assuming that a number of defect areas have been located in a batch of yarn surface images, the image processing tool will divide these areas into multiple independent units, such as dividing an image into 10 small areas, each representing a different part of the yarn surface. For each area, feature parameters such as color depth, texture continuity, etc. are extracted to form a set of regional feature distributions. The purpose of this process is to gain a more detailed understanding of the surface characteristics of each area, laying the foundation for subsequent analysis. For the set of regional feature distributions, the data comparison tool will quantify the defect density of each area. Assuming that the preset defect density threshold is 5 defect points per square centimeter, if the defect density of a certain area reaches 7 defect points per square centimeter, it will be marked as a high-density area. Finally, assuming that 3 of the 10 areas are marked as high-density areas, a set of high-density areas is formed. This quantitative processing helps to quickly identify the parts of the yarn surface where problems are more concentrated, providing key objects for subsequent processing. On the basis of the set of high-density areas, the feature extraction tool will further calculate the feature deviation of each area. Assuming that the feature deviation is mainly measured by color distribution and texture uniformity, and the preset deviation threshold is 10%, while the color deviation of a certain area reaches 15%, it will be marked as a deviation abnormal area. Finally, 2 of the 3 high-density areas are marked as deviation abnormal areas, forming a set of deviation abnormal areas. This way can more accurately screen out key areas that may affect the quality of the yarn.

[0067] Further, obtaining the final detection analysis result includes:

[0068] According to the multi-dimensional evaluation data, the yarn uniformity distribution is divided into areas, at least one feature parameter is extracted for each area, the regional distribution data set is obtained, and a preliminary distribution feature set is obtained;

[0069] Through the distribution feature set, color mapping processing is performed on the defect high-incidence area, the defect density level is determined for the color mapping result, and an intuitive distribution presentation graph is obtained;

[0070] If the color value of some areas in the distribution presentation graph is higher than the preset threshold, spatial positioning processing is performed to obtain specific area labeling information, and the position point of focus is determined;

[0071] According to the area labeling information and the distribution presentation graph, the position information and the visual report content are fused to generate a final detection result containing defect high-incidence area labeling.

[0072] Specifically, in the business scenario of yarn surface quality detection, the processing and analysis of multi-dimensional evaluation data are particularly important. When dividing the yarn uniformity distribution based on image processing tools, the entire yarn surface image can be segmented into multiple small regions, such as 8 equal-area units, each representing a different part of the yarn surface. Feature parameters such as surface roughness and color difference are extracted for each region to form a regional distribution dataset. This division and extraction method can help capture the local characteristics of the yarn surface in more detail and lay the foundation for subsequent analysis. By using the above distribution feature set, when color mapping the defect-prone areas using a heat map generation tool, the color can be gradually changed from light to dark according to the defect density. Assuming that a certain area has a large number of defect points, the color mapping is deep red, while another area with fewer defects is mapped to light yellow, and finally an intuitive distribution presentation graph is generated. This color mapping method directly reflects the concentration of defect distribution, making it easy to quickly identify problem areas. The distribution of yarn surface defects directly affects the quality of the final textile product. If high-risk areas can be identified in time through heat maps and labeling information, production personnel can adjust the process accordingly to avoid further defects. This systematic analysis and presentation method not only improves detection efficiency but also provides a reliable basis for quality control, helping to reduce the rate of defective products and ensure the overall quality of the yarn. In one possible implementation, when determining the defect density level based on the color mapping result, the density can be divided into high, medium, and low levels, corresponding to different color intervals, such as deep red for high density, yellow for medium density, and green for low density. This grading method facilitates quick judgment of the severity of defects in each area and provides a reference for subsequent processing. At the same time, this intuitive grading presentation also makes it easy for non-professionals to understand the detection results, enhancing the readability of the report.

[0073] The application discloses a yarn uniformity analysis method based on image processing. To solve the problem of yarn image acquisition under complex lighting conditions, an initial image is obtained through a multi-light source adaptive adjustment system, and lighting balance processing is performed. Then, wavelet transform and convolutional neural network are used to extract yarn surface features, perform local magnification analysis and edge detection on abnormal areas, and determine defect candidate areas. Through clustering analysis and feature similarity calculation, non-defect areas are removed, and finally accurate defect positioning results are obtained. The application also uses a weighted average algorithm to calculate uniformity evaluation scores and generate a visual analysis report to intuitively present yarn uniformity distribution and defect positions. The application can effectively solve the problems of yarn image acquisition and defect detection under complex lighting conditions, improve the accuracy and efficiency of yarn uniformity detection, and provide important technical support for textile quality control.

[0074] The above merely provides the preferred embodiments of the present application, and the protection scope of the present application is not limited thereto, and any changes or substitutions within the technical scope disclosed by the present application should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A yarn uniformity analysis method based on image processing, characterized in that, include: Acquire initial yarn image data, and obtain a first yarn image based on the initial yarn image data; Based on the brightness distribution characteristics of the first yarn image, local contrast enhancement processing is performed on the uneven lighting area, and grayscale normalization is performed on the shadow interference area of ​​the enhanced image to obtain a second yarn image with uniform brightness. To address the differences in texture features in the second yarn image, wavelet transform image decomposition technology is used to divide the image into multiple frequency sub-bands, extract texture detail information in the high-frequency sub-bands, and obtain a preliminary feature distribution map of the yarn surface. Using the preliminary feature distribution map, deep learning analysis is performed on the surface features of the yarn, and adaptive feature weighting is applied to the analyzed complex texture areas to obtain a more accurate third feature distribution map. The regions in the third feature distribution map whose feature values ​​exceed a preset abnormal threshold are subjected to local magnification analysis, and the boundaries of subtle defects are further extracted by combining edge detection algorithms to obtain the final set of defect candidate regions. Based on the set of defect candidate regions, the candidate regions are grouped, and a feature similarity index is calculated for each group. If the similarity index is lower than a preset discrimination threshold, the non-defect region is eliminated, and the final defect location result is determined. Based on the final defect location results and the overall characteristic distribution of the yarn, a uniformity evaluation score is calculated. The defect density and characteristic deviation in different areas are comprehensively quantified to obtain multi-dimensional evaluation data of yarn uniformity. Based on the multi-dimensional evaluation data, a visual analysis report is generated to intuitively present the yarn uniformity distribution, mark the specific location information of high-defect areas, and obtain the final detection and analysis results.

2. The yarn uniformity analysis method based on image processing as described in claim 1, characterized in that, Obtaining initial yarn image data includes: Based on the multi-source adaptive adjustment mechanism, and considering the yarn characteristics under complex lighting conditions, a multi-angle illumination method is adopted. Environmental interference data is obtained through a preset light intensity sensor, and the output intensity of the light source is adjusted in real time by a light intensity dynamic module to obtain the initial yarn image data.

3. The yarn uniformity analysis method based on image processing as described in claim 1, characterized in that, Obtaining the first yarn image based on the initial yarn image data includes: Using the initial yarn image data, an image acquisition device is used to acquire an initial image under multi-angle illumination conditions. If uneven illumination distribution is detected to exceed a preset threshold due to environmental interference factors, the light source output is further adjusted through a light intensity dynamic module to obtain the first yarn image.

4. The yarn uniformity analysis method based on image processing as described in claim 1, characterized in that, Obtaining a second yarn image with uniform brightness includes: Based on the brightness distribution characteristics of the first yarn image, local contrast enhancement processing is performed on the uneven lighting area, and the intermediate yarn image after preliminary adjustment is obtained by adjusting the pixel value distribution range. For the shadow interference area in the intermediate yarn image, the gray value of the local area is standardized. If the gray value of the shadow interference area is detected to deviate from the preset threshold, the shadow interference area is adjusted a second time to obtain a gray-scale balanced transition yarn image. By detecting the overall brightness distribution of the transition yarn image, the areas where there is still uneven illumination are refined to obtain an optimized yarn image with a more uniform brightness distribution. Based on the brightness characteristics of the optimized yarn image, the grayscale value and contrast of the optimized yarn image are finally detected. If it is found that a local area still does not meet the preset threshold, fine-tuning is performed using a local adjustment tool to obtain a second yarn image with balanced brightness.

5. The yarn uniformity analysis method based on image processing as described in claim 1, characterized in that, Obtaining a preliminary feature distribution map of the yarn surface includes: Based on the texture characteristics of the second yarn image, the image is segmented into multiple levels, and different frequency components are separated to obtain high-frequency sub-band data and obtain a frequency layered image after preliminary decomposition. For the high-frequency subband in the frequency layered image, the texture details are separated. If the signal strength of the high-frequency subband exceeds a preset threshold, enhancement processing is performed to determine the set of detailed features on the yarn surface. By performing distribution analysis on the set of detailed features, the extracted texture details are correlated and matched with the surface characteristics of the yarn to obtain the distribution pattern corresponding to the surface characteristics and obtain a preliminary feature distribution map of the yarn surface.

6. The yarn uniformity analysis method based on image processing as described in claim 1, characterized in that, Obtaining a more accurate third feature distribution map includes: Based on the preliminary feature distribution map, the complex texture areas on the yarn surface are divided into regions to obtain surface detail data of at least one target region, resulting in a set of region details after division. For the set of details in the region, the texture distribution of each region is separated. If the texture complexity of the region is detected to exceed a preset threshold, the region is adaptively adjusted to determine the weighted feature set. The weighted feature set is used to process the feature matching of the yarn surface to obtain more accurate surface detail distribution data and obtain optimized feature mapping results. Based on the optimized feature mapping results, the distribution map of the yarn surface is updated, and the surface details of different regions are integrated to obtain a more accurate third feature distribution map.

7. The yarn uniformity analysis method based on image processing as described in claim 1, characterized in that, The final set of candidate defect regions includes: Based on the third feature distribution map, the target area exceeding the preset abnormal threshold is locally magnified to obtain magnified surface detail data and obtain a magnified detail view. Based on the magnified detailed view, boundary recognition processing is performed on minor defects. If the clarity of the detected defect boundary is lower than a preset threshold, a secondary sharpening process is performed to determine the range of clear defect boundaries. Based on the clearly defined defect boundary range, the defect region is divided, and at least one independent candidate region is extracted from it to obtain a preliminary set of defect candidate regions. Based on the preliminary set of candidate defect regions, the surface details and distribution map of each candidate region are matched. If the matching degree is lower than a preset threshold, details are extracted and supplemented to obtain the final set of candidate defect regions.

8. The yarn uniformity analysis method based on image processing as described in claim 1, characterized in that, The final defect location results include: Based on the final set of candidate defect regions, the candidate regions are initially grouped. For each region in the group, its surface feature data is obtained, and at least one feature vector is extracted from it to obtain the initial grouping result. Based on the preliminary grouping results, the similarity of the feature vectors within the group is calculated. If the calculated similarity index is lower than the preset discrimination threshold, the corresponding region is marked as a non-defect region and removed to determine the filtered region set. By comparing the surface detail depth of the filtered region set, detailed feature data is obtained, and it is determined whether the region meets the target defect standard, thus obtaining a candidate region set that meets the standard. Based on the set of candidate regions that meet the criteria, the regions are subjected to boundary confirmation processing. If the boundary clarity reaches a preset threshold, the region is confirmed as the final defect location, and the final defect location result is obtained.

9. The yarn uniformity analysis method based on image processing as described in claim 1, characterized in that, The multi-dimensional evaluation data for obtaining yarn evenness includes: Based on the final defect location result, the yarn surface is divided into regions, and at least one feature parameter is extracted for each region to obtain a set of region feature distributions. The defect density of each region is quantified using the set of regional feature distributions. If the defect density is higher than a preset threshold, it is marked as a high-density region, and a set of high-density regions is determined. Based on the set of high-density regions, feature extraction tools are used to calculate the feature deviation of each region and obtain feature deviation data. If the feature deviation exceeds a preset threshold, it is marked as a region with abnormal deviation, and a set of regions with abnormal deviation is obtained. By combining the set of abnormal deviation regions with the set of regional feature distributions, a weighted average calculation is used to comprehensively process the yarn characteristics and defect density of the regions, thereby obtaining multi-dimensional evaluation data of yarn uniformity.

10. The yarn uniformity analysis method based on image processing as described in claim 1, characterized in that, The final test and analysis results include: Based on the multi-dimensional evaluation data, the yarn uniformity distribution is divided into regions, and at least one feature parameter is extracted for each region to obtain the regional distribution dataset and a preliminary distribution feature set. Using the aforementioned distribution feature set, color mapping is performed on areas with high defect incidence, and the defect density level is determined based on the color mapping results to obtain an intuitive distribution map. If the color value of certain areas in the distribution map is higher than a preset threshold, spatial positioning processing is performed to obtain specific area labeling information and determine the key locations to focus on. Based on the region labeling information and the distribution map, the location information and the visualization report content are fused together to generate a final detection result that includes labels of areas with high defect incidence.

Images