PVC floor surface coating defect detection method and system

Abstract

The present invention relates to the field of surface quality inspection technology, and in particular to a method and system for detecting defects in the surface coating of PVC flooring. The method comprises: dividing the surface of the coating to be inspected into regions, obtaining regional scan images, and performing image optical compensation on the regional scan images; analyzing the high-dimensional image features of each regional scan image, performing a difference analysis between the high-dimensional image features and reference features, and selecting surface defect regions; establishing a high-dimensional defect feature analysis model, analyzing and generating defect location information and defect geometric parameters, and determining the defect type of the surface defect region based on the defect location information and the defect geometric parameters; and quantitatively evaluating the sensitivity weight of the defect to the surface performance of the coating to be inspected based on the defect location information and the defect geometric parameters, and outputting an overall surface performance impact analysis result. The present invention effectively solves the problem of insufficient recognition capability for small-sized and complex-shaped defects.

Description

Technical Field

[0001] The present invention relates to the technical field of surface quality detection, and in particular to a method and system for detecting defects in the surface coating of a PVC floor. Background Art

[0002] In the production of PVC flooring and similar coated surface materials, coating quality directly impacts the product's performance and aesthetic appearance. However, due to uneven application, material defects, or external contamination during the production process, coating surfaces are prone to various defects, such as bubbles, cracks, and uneven areas. These defects can lead to reduced mechanical properties, durability, and appearance. Detecting and analyzing coating surface defects has become a critical step in quality control.

[0003] Existing technologies typically employ detection methods based on two-dimensional image acquisition, identifying coating defects through defect classification techniques based on fixed image recognition rules. However, these methods suffer from deficiencies in detection accuracy, adaptability, and performance evaluation. Their ability to identify small or complex defects is limited, and defects with irregular surface boundaries are particularly prone to false detection or missed detection. Furthermore, these methods typically rely on fixed rules or thresholds and lack dynamic adjustment capabilities. Furthermore, most existing technologies focus solely on the presence of defects, lacking quantitative analysis of their impact on the overall performance of the coating, and are unable to fully assess the actual impact of defects on the durability and quality of the coating.

[0004] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present disclosure and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art known to those skilled in the art. Summary of the Invention

[0005] The present invention provides a method and system for detecting defects in the surface coating of PVC flooring, which can effectively solve the problems in the background technology.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] A method for detecting defects in a PVC floor surface coating, the method comprising:

[0008] Dividing the surface of the coating to be tested into regions, acquiring regional scanning images, and performing image optical compensation on the regional scanning images;

[0009] Analyzing high-dimensional image features of the scanned image of each region, performing difference analysis between the high-dimensional image features and the reference features, and selecting the surface defect region;

[0010] Establishing a high-dimensional defect feature analysis model, analyzing and generating defect location information and defect geometric parameters, and determining the defect type of the surface defect area based on the defect location information and the defect geometric parameters;

[0011] According to the defect location information and the defect geometric parameters, a quantitative evaluation is performed in combination with the sensitivity weight of the defect to the surface performance of the coating to be tested, and an overall surface performance impact analysis result is output.

[0012] Furthermore, surface defect areas are selected, including:

[0013] Performing multi-scale feature decomposition on the area scan image to extract the high-dimensional image features including texture features, edge features, and brightness distribution features;

[0014] Comparing the high-dimensional image features with the reference features, training and generating a difference analysis algorithm, calculating difference indicators based on the difference analysis algorithm and identifying potential defect areas;

[0015] Performing regional connectivity analysis on the potential defect region based on the reference features to generate a preliminary surface defect region;

[0016] The boundary of the preliminary surface defect area is finely adjusted based on the difference analysis algorithm to generate the surface defect area.

[0017] Furthermore, performing regional connectivity analysis on the potential defect region based on the benchmark features includes:

[0018] Performing neighborhood connectivity calculation on the pixels within the potential defect area based on pixel grayscale values ​​and geometric structure continuity to screen out defective pixel groups that meet the connectivity requirements;

[0019] Performing cluster analysis on the defective pixel group and further grouping them to generate multiple connected sub-regions and calculating texture consistency of the connected sub-regions;

[0020] Performing connectivity analysis on the morphological parameters of the connected sub-regions, and eliminating pseudo-defect regions where the texture consistency is lower than a set value based on the reference feature comparison;

[0021] According to the results of the connectivity analysis, a preliminary surface defect region is generated.

[0022] Furthermore, calculating the texture consistency of the connected sub-regions includes:

[0023] Analyzing the connected sub-regions and obtaining texture direction features and texture frequency features of the connected sub-regions;

[0024] Calculating and quantifying the consistency of the texture direction within the connected sub-region based on the texture direction feature to generate a direction standard deviation;

[0025] Calculating the complexity of texture details of the connected sub-region based on the texture frequency feature to generate a texture spectrum distribution deviation;

[0026] The direction standard deviation and the texture spectrum distribution deviation are weighted and comprehensively calculated to comprehensively generate the texture consistency of the connected sub-region.

[0027] Furthermore, determining the defect type of the surface defect area includes:

[0028] Generating the high-dimensional defect feature analysis model according to the high-dimensional image features and the reference feature training;

[0029] Scanning the surface defect area pixel by pixel to generate a pixel scanning result;

[0030] Calculate the center point coordinates and area boundary information of the defect position based on the pixel scanning result to generate the defect location information;

[0031] Calculating the area, perimeter, and shape regularity of the defect based on the pixel scanning result and the defect location information to generate the defect geometric parameters;

[0032] The defect location information and the defect geometric parameters are associated with each other based on the high-dimensional defect feature analysis model to determine the defect type of the surface defect area.

[0033] Furthermore, the center point coordinates and area boundary information of the defect position are calculated based on the pixel scanning result, including:

[0034] Performing pixel classification screening on the pixel scanning results, marking pixels that meet a preset difference threshold as defective pixels, and generating a defective pixel set;

[0035] Performing connectivity analysis on the defective pixel set to screen out defective areas that meet connectivity rules;

[0036] Calculating a minimum circumscribed rectangle of the surface defect area according to pixel distribution of the surface defect area;

[0037] Calculating the center point coordinates of the defect position based on the geometric center of the minimum circumscribed rectangle;

[0038] Boundary information of the minimum circumscribed rectangle is obtained to generate region boundary information of the surface defect region.

[0039] Furthermore, a quantitative evaluation is performed on the sensitivity weight of the surface performance of the coating to be tested in combination with the defects, including:

[0040] Determining the relative position distribution of the surface defect area in the total area of ​​the coating surface to be tested according to the defect location information, and calculating the area position weight;

[0041] Calculating a geometric influence factor based on the defect geometric parameters, the geometric influence factor including the influence of the area ratio on the overall integrity and the shape regularity deviation value;

[0042] Calculating a sensitivity weight by combining the regional position weight and the geometric influence factor, wherein the sensitivity weight is used to quantify the potential impact of the defect on the coating performance;

[0043] For each of the surface defect areas, calculating a performance impact value of the single surface defect area according to the sensitivity weight and the geometric impact factor;

[0044] The performance impact values ​​of all the surface defect areas are cumulatively calculated to generate an overall surface performance impact analysis result.

[0045] Furthermore, the geometric impact factor is calculated, including:

[0046] Obtaining the area and perimeter of the surface defect region, and calculating the area ratio and boundary complexity of the defect region;

[0047] Calculating a shape regularity deviation value of the defective area based on the area ratio of the defective area, wherein the shape regularity deviation value is determined based on a ratio of an actual perimeter of the defective area to a theoretical minimum perimeter;

[0048] Calculating a boundary complexity coefficient according to the boundary complexity of the surface defect area and based on the detail change frequency and the number of mutation points of the defect boundary;

[0049] The geometric impact factor is generated by combining the shape regularity deviation value and the boundary complexity coefficient.

[0050] A PVC floor surface coating defect detection system, the system comprising:

[0051] The area scanning module divides the coating surface to be tested into areas, obtains area scanning images, and performs image optical compensation on the area scanning images;

[0052] The defect selection module analyzes the high-dimensional image features of the scanned image of each area, performs difference analysis between the high-dimensional image features and the reference features, and selects the surface defect areas;

[0053] Type judgment module: establishes a high-dimensional defect feature analysis model, analyzes and generates defect location information and defect geometric parameters, and judges the defect type of the surface defect area based on the defect location information and defect geometric parameters;

[0054] The performance evaluation module quantitatively evaluates the sensitivity weight of the defect to the surface performance of the coating to be tested based on the defect location information and defect geometric parameters, and outputs the overall surface performance impact analysis results.

[0055] Furthermore, the defect selection module includes:

[0056] Feature decomposition unit, which performs multi-scale feature decomposition on the area scan image and extracts high-dimensional image features including texture features, edge features and brightness distribution features;

[0057] The potential marking unit compares the high-dimensional image features with the reference features to train and generate a difference analysis algorithm. Based on the difference analysis algorithm, the difference index is calculated and the potential defect area is identified.

[0058] Connectivity analysis unit, which performs regional connectivity analysis on potential defect areas based on benchmark features and generates preliminary surface defect areas;

[0059] The optimization and adjustment unit makes fine adjustments to the boundaries of the preliminary surface defect area based on the difference analysis algorithm to generate the surface defect area.

[0060] The technical solution of the present invention can achieve the following technical effects:

[0061] It can achieve high-precision detection of complex defects on the coating surface, and combine the location information and geometric parameters of the defects to quantitatively evaluate the impact of defects on the overall performance of the coating, adapt to defects of different types and distributions, and improve the accuracy and adaptability of detection.

[0062] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0064] Figure 1 The figure is a flow chart of a method for detecting defects in the surface coating of PVC flooring;

[0065] Figure 2 Schematic diagram of the process for generating surface defect areas;

[0066] Figure 3 Schematic diagram of the structure for determining the type of surface defects;

[0067] Figure 4 Schematic diagram of the structure for quantitative evaluation of the surface sensitivity weight of the coating to be tested. DETAILED DESCRIPTION

[0068] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0069] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0070] Embodiment 1;

[0071] like Figure 1 As shown, the present application provides a method for detecting defects in the surface coating of PVC flooring, the method comprising:

[0072] Divide the coating surface to be tested into regions, obtain regional scanning images, and perform image optical compensation on the regional scanning images;

[0073] Analyze the high-dimensional image features of the scanned image of each area, perform difference analysis between the high-dimensional image features and the reference features, and select the surface defect area;

[0074] Establish a high-dimensional defect feature analysis model to analyze and generate defect location information and defect geometric parameters, and determine the defect type of the surface defect area based on the defect location information and defect geometric parameters;

[0075] Based on the defect location information and defect geometric parameters, the sensitivity weight of the defect to the surface performance of the coating to be tested is quantitatively evaluated, and the overall surface performance impact analysis results are output.

[0076] Specifically, the surface of the PVC floor coating to be tested is divided into regions. Specifically, the surface to be tested can be divided into multiple detection regions according to a preset grid size to ensure that each region evenly covers the coating surface, and an image of each region is collected by a high-resolution industrial camera or laser scanning equipment to generate a regional scan image. The collected regional scan image is optically compensated. The compensation method includes adjusting the brightness and contrast of the image to eliminate the brightness difference caused by uneven ambient lighting, correcting the color distortion of the image based on the reference color model to ensure the true reflection of the image, and using algorithms such as Gaussian filtering or median filtering to remove possible background noise or random noise in the image; the regional scan image after optical compensation is High-dimensional image features, such as geometric, texture, and color features, are extracted. These features are then compared with baseline features (pre-constructed from reference features of normal coating areas) for difference analysis. Algorithms such as Euclidean distance or cosine similarity are used to calculate feature differences and select surface defect areas. A high-dimensional defect feature analysis model is constructed and trained using a support vector machine (SVM) or deep learning model (such as a convolutional neural network (CNN)) based on a large amount of labeled defect data. The defect areas selected through the difference analysis are then fed into the high-dimensional defect feature analysis model to generate defect location information and defect geometric parameters. Based on this defect location information and defect geometric parameters, the analysis model classifies and determines defect types, such as bubbles, cracks, and uneven coatings. Weights are assigned based on defect location (e.g., edge or center), with edge areas receiving greater weights than center areas. Weights are adjusted based on defect geometric parameters (e.g., area, regularity, and boundary complexity), with more significant defects receiving higher weights. The performance impact of each surface defect area is calculated by combining its sensitivity weight and the defect geometric parameters. The performance impact values ​​of all defect areas are then accumulated to generate a performance impact analysis result for the entire coating surface, outputting a quantitative assessment of coating performance.

[0077] Through the technical solution of the present invention, high-precision detection of complex defects on the coating surface can be achieved, and the impact of defects on the overall performance of the coating can be quantitatively evaluated by combining defect location information and geometric parameters, adapting to defects of different types and distributions, and improving the accuracy and adaptability of detection.

[0078] Further, if Figure 2 As shown, select the surface defect areas, including:

[0079] Perform multi-scale feature decomposition on the area scan image to extract high-dimensional image features including texture features, edge features and brightness distribution features;

[0080] Compare high-dimensional image features with baseline features and train them to generate a difference analysis algorithm. Based on the difference analysis algorithm, calculate the difference index and identify potential defect areas.

[0081] Perform regional connectivity analysis on potential defect areas based on benchmark features to generate preliminary surface defect areas;

[0082] The boundary of the preliminary surface defect area is finely adjusted based on the difference analysis algorithm to generate the surface defect area.

[0083] As a preferred embodiment of the above embodiment, multi-scale feature decomposition is performed on the regional scan image to extract high-dimensional image features, including texture features, edge features and brightness distribution features. The image is decomposed by methods such as wavelet transform or Gaussian pyramid to obtain frequency components of different resolutions. The gray-level co-occurrence matrix is ​​used to extract information such as texture direction, consistency and contrast. The Canny edge detection or Sobel operator is used to extract the gradient direction and intensity distribution of the edge. At the same time, the brightness abnormality area is identified through the brightness histogram and distribution curve, so as to fully obtain the surface characteristics of the coating; the extracted high-dimensional image features are compared with the benchmark features to generate a difference analysis algorithm. The benchmark features are constructed from the defect-free coating surface image, including texture, edge and brightness characteristics. The comparison training is performed through a supervised learning method (such as a support vector machine or a random forest). The difference analysis algorithm is used to calculate the difference index of the features. (such as Euclidean distance or cosine similarity), use thresholds to screen out image areas with significant differences, preliminarily identify potential defect areas, and further improve regional integrity through regional merging; perform regional connectivity analysis on potential defect areas, use 4-neighborhood or 8-neighborhood connectivity algorithms to screen out pixel groups that meet connectivity rules, eliminate pseudo-defect areas with too small an area or irregular shape, generate preliminary surface defect areas, and eliminate noise interference through connectivity analysis; based on the difference analysis algorithm, fine-tune the boundaries of the preliminary surface defect areas, use active contour models or improved edge detection methods to extract the boundaries of the defect areas, and combine the comparison results of high-dimensional image features with benchmark features to smooth or locally correct the boundaries to optimize boundary accuracy. Finally, the optimized defect areas are superimposed and confirmed with the original image to ensure the boundary clarity and accuracy of the detection results.

[0084] Furthermore, regional connectivity analysis of potential defect areas is performed based on the benchmark features, including:

[0085] The neighborhood connectivity of pixels in the potential defect area is calculated based on the pixel grayscale value and geometric structure continuity, and defective pixel groups that meet the connectivity are screened out;

[0086] Perform cluster analysis on the defective pixel groups and further group them to generate multiple connected sub-regions and calculate the texture consistency of the connected sub-regions;

[0087] Connectivity analysis is performed on the morphological parameters of connected sub-regions, and pseudo-defect regions with texture consistency lower than the set value are eliminated based on the reference feature comparison;

[0088] Based on the results of the connectivity analysis, a preliminary surface defect region is generated.

[0089] As a preferred embodiment of the above, the neighborhood connectivity calculation is performed on the pixels in the potential defect area, and the pixel groups that meet the connectivity rules are screened out through grayscale value and geometric structure continuity analysis. The specific method includes calculating the grayscale difference between the pixel and the surrounding pixels, setting a threshold to determine whether they belong to the same area, and using a 4-neighborhood or 8-neighborhood algorithm to further analyze the spatial connection relationship of the pixels to ensure the geometric continuity of the area; clustering analysis and grouping are performed on the screened defective pixel groups to generate multiple connected sub-regions, and their texture consistency is calculated. The defective pixel groups are divided into several connected sub-regions through a density clustering algorithm or a distance-based K-means algorithm to ensure that each region is spatially complete and independent. , extract the texture features (such as direction, contrast and consistency) of each connected sub-region, and calculate the texture consistency value by counting the internal pixel gradient changes to quantify the texture characteristics of each sub-region; further perform morphological parameter analysis on the generated connected sub-regions, and combine the benchmark features to eliminate pseudo-defect areas with low texture consistency, extract morphological parameters such as area, perimeter, shape regularity and boundary complexity of the sub-regions, compare the morphological parameters with the normal area parameters in the benchmark features, eliminate areas with large deviations from the benchmark features, and according to the texture consistency threshold, eliminate areas with consistency lower than the set value, thereby removing pseudo-defects; according to the results of the connectivity analysis, integrate the screened connected sub-regions to generate a preliminary surface defect area.

[0090] Furthermore, the texture consistency of the connected sub-region is calculated, including:

[0091] Analyze the connected sub-regions and obtain the texture direction features and texture frequency features of the connected sub-regions;

[0092] The consistency of texture direction within the connected sub-region is calculated and quantified based on texture direction features to generate the direction standard deviation;

[0093] Calculate the complexity of texture details in connected sub-regions based on texture frequency features and generate texture spectrum distribution deviation;

[0094] The direction standard deviation and texture spectrum distribution deviation are weighted and comprehensively calculated to generate the texture consistency of the connected sub-region.

[0095] As a preferred embodiment of the above embodiment, texture direction features and texture frequency features are extracted from the connected sub-region. The texture direction features are used to calculate the texture direction distribution through the gray level co-occurrence matrix (GLCM), and the Sobel operator or the histogram of directional gradients (HOG) is used to analyze the directionality of the edge in the region. The texture frequency features are used to analyze the frequency distribution of the region through the fast Fourier transform (FFT) or wavelet transform, and the high-frequency and low-frequency energy ratios of the texture are extracted as the core indicators of the frequency features. Based on the extracted texture direction features, the directional consistency within the connected sub-region is calculated. By counting the directional gradient value of each pixel, the distribution of the directional gradient in the region is calculated and the directional standard deviation is generated. The directional standard deviation reflects the The consistency of texture direction, the smaller the standard deviation, the more consistent the texture direction, and vice versa; the complexity of texture details in the connected sub-region is calculated based on the texture frequency characteristics, the frequency distribution map of the connected sub-region is obtained through frequency domain conversion, the center value and energy distribution curve of the spectrum are analyzed, and the deviation value of the spectrum distribution from the reference spectrum is calculated; the direction standard deviation and the texture spectrum distribution deviation are weighted and integrated to generate the texture consistency value of the connected sub-region, the direction consistency and spectrum deviation are unified to the same dimensional range through normalization processing, and the linear weighted model is used to comprehensively calculate the consistency value. The importance of direction consistency and spectrum consistency can be flexibly adjusted through weight factors, and finally the texture consistency of the connected sub-region is output.

[0096] Further, if Figure 3 As shown, the defect types of the surface defect area are judged, including:

[0097] Generate a high-dimensional defect feature analysis model based on high-dimensional image features and benchmark feature training;

[0098] Scan the surface defect area pixel by pixel to generate pixel scanning results;

[0099] Calculate the center point coordinates and area boundary information of the defect position based on the pixel scanning results to generate defect location information;

[0100] Based on the pixel scanning results and defect location information, the area, perimeter and shape regularity of the defect are calculated to generate the defect geometric parameters;

[0101] The defect location information and defect geometric parameters are associated based on the high-dimensional defect feature analysis model to determine the defect type of the surface defect area.

[0102] As a preferred embodiment of the above, a classification model is constructed based on high-dimensional image features and reference features through training of a high-dimensional defect feature analysis model, and texture features (such as direction, consistency), geometric features (such as area, shape) and brightness features are extracted from the defect-free coating area to generate a reference feature library. In combination with the labeled defect sample data (such as bubbles, cracks, uneven coating, etc.), high-dimensional image features of the samples are extracted, and training is performed using a classification algorithm (such as a support vector machine or a convolutional neural network) to generate a high-dimensional defect feature analysis model; the surface defect area is scanned pixel by pixel, the pixel scanning results are extracted and defect location information is generated, and the grayscale value, position coordinates and neighborhood connectivity information of each pixel point in the defect area are recorded pixel by pixel to form a pixel scanning matrix. Based on the pixel distribution, the centroid coordinates of the defect area are calculated as the center point position, and the edge detection The measurement algorithm (such as Canny or Sobel operator) extracts boundary contour information to fully describe the spatial positioning characteristics of the defect area; based on the pixel scanning results and defect positioning information, the defect geometric parameters are further calculated. The area of ​​the defect area is calculated by counting the total number of pixels in the defect area, the perimeter is calculated using the boundary contour information, and the regularity of the defect shape is calculated by the ratio of area to perimeter. The regularity index can reflect the deviation between the defect shape and the ideal geometric shape; the defect type is classified through a high-dimensional defect feature analysis model, and the defect positioning information (center point coordinates, boundary contour) is associated with the geometric parameters (area, perimeter, regularity) to generate a complete defect feature vector. The feature vector is input into the classification model, and the defect is identified in combination with the feature classification rules generated by training, and finally the defect type label (such as bubbles, cracks, uneven coating, etc.) is output.

[0103] Furthermore, the center point coordinates and area boundary information of the defect location are calculated based on the pixel scanning results, including:

[0104] Perform pixel classification screening on the pixel scanning results, mark pixels that meet the preset difference threshold as defective pixels, and generate a defective pixel set;

[0105] Perform connectivity analysis on the defective pixel set and filter out defective areas that meet the connectivity rules;

[0106] Calculate the minimum circumscribed rectangle of the surface defect area according to the pixel distribution of the surface defect area;

[0107] Calculate the center point coordinates of the defect location based on the geometric center of the minimum circumscribed rectangle;

[0108] The boundary information of the minimum circumscribed rectangle is obtained to generate the region boundary information of the surface defect region.

[0109] As a preferred embodiment of the above, a defective pixel set is extracted from the pixel scanning result through pixel classification screening. The specific method is to set a grayscale difference threshold based on the benchmark feature of the defect-free area, compare the grayscale value in the scanning result pixel by pixel, mark the pixels exceeding the threshold as defective pixels, and gather all the marked defective pixels to form a defective pixel set. The defective pixel set is processed by connectivity analysis, and the spatial connection relationship between the pixels is judged by using the 4-neighborhood or 8-neighborhood rule. The pixel group with coherence is screened out as the defective area, and the area and shape characteristics of the area are counted to eliminate pseudo defects with too small an area or abnormal shape. The minimum bounding rectangle of the defective area generated based on the connectivity analysis is further calculated. Boundary pixel points of the defect area are extracted to form a boundary point set. The minimum bounding box algorithm is used to calculate the minimum enclosing rectangle covering all boundary points, and the vertex coordinates and side length information of the rectangle are output. The center point coordinates of the defect position are calculated based on the minimum enclosing rectangle. Through the vertex coordinates of the enclosing rectangle, its geometric center is calculated as the center point of the defect area, specifically the midpoint position of the left and right boundaries and the upper and lower boundaries of the rectangle, to clarify the position of the defect on the coating surface; the regional boundary information of the defect is generated using the minimum enclosing rectangle, and the vertex coordinates, side length and angle information of the rectangle are extracted, and these data are stored as the regional boundary information of the defect area. The regional boundary information can be superimposed on the original image for visualization, intuitively showing the shape and position of the defect.

[0110] Further, if Figure 4 As shown, the sensitivity weights of defects to the surface properties of the coating to be tested are quantitatively evaluated, including:

[0111] Determine the relative position distribution of the surface defect area in the total area of ​​the coating surface to be tested based on the defect location information, and calculate the regional position weight;

[0112] Based on the defect geometric parameters, the geometric influence factor is calculated. The geometric influence factor includes the influence of the area ratio on the overall integrity and the shape regularity deviation value;

[0113] The sensitivity weight is calculated by combining the regional position weight and geometric influence factor. The sensitivity weight is used to quantify the potential impact of defects on coating performance.

[0114] For each surface defect area, the performance impact value of a single surface defect area is calculated based on the sensitivity weight and geometric impact factor;

[0115] The performance impact values ​​of all surface defect areas are cumulatively calculated to generate the overall surface performance impact analysis results.

[0116] As a preferred embodiment of the above, the relative position distribution of the surface defect area in the total area of ​​the coating surface is calculated through the defect positioning information, and the regional position weight is assigned, that is, based on the coordinates of the center point of the defect, it is mapped to the global coordinate system of the coating, and the weight is assigned according to the importance of the position. For example, the edge area is more sensitive to the coating performance, so the weight is higher, while the center area has a lower weight. The calculated regional position weight is used to quantify the influence of the position on the coating performance and provide an important basis for defect analysis; the geometric influence factor is calculated based on the geometric parameters of the defect, and the coverage of the defect on the integrity of the coating is directly reflected by the statistical ratio of the defect area to the total area of ​​the coating. The abnormality of the defect shape is quantified by the shape regularity deviation value. The regularity deviation value is calculated based on the ratio of the area to the perimeter. The larger the deviation, the more irregular the defect shape. , the area ratio and regularity deviation value are combined to generate a geometric influence factor to further reflect the influence of geometric characteristics on performance; the sensitivity weight of the defect is calculated by combining the regional position weight and the geometric influence factor, and the position weight and the geometric influence factor are combined through a weighted formula. The weight factor can adjust the relative importance of the position and geometric characteristics according to the application scenario. The sensitivity weight is normalized to represent the potential impact of the defect area on the coating performance; based on the sensitivity weight and the geometric influence factor, the performance impact value of each defect area is calculated. The performance impact value is the weighted product of the two to generate a performance impact metric for a single defect area. The performance impact values ​​of all surface defect areas are cumulatively added to generate the performance impact analysis results of the overall coating surface. The results are output in numerical or graphical form to intuitively show the extent to which the coating surface performance is affected.

[0117] Further, the geometric impact factor is calculated, including:

[0118] Obtain the area and perimeter of the surface defect area, and calculate the area ratio and boundary complexity of the defect area;

[0119] Based on the area ratio of the defect area, the shape regularity deviation value of the defect area is calculated. The shape regularity deviation value is determined based on the ratio of the actual perimeter of the defect area to the theoretical minimum perimeter;

[0120] According to the boundary complexity of the surface defect area, the boundary complexity coefficient is calculated based on the detail change frequency and the number of mutation points of the defect boundary;

[0121] The geometric impact factor is generated by combining the shape regularity deviation value and the boundary complexity coefficient.

[0122] As a preferred embodiment of the above, the area and perimeter of the defective area are calculated to obtain its basic geometric parameters and evaluate the regional characteristics, that is, the number of pixels in the defective area is counted and the ratio is calculated to the total area of ​​the coating to obtain the area ratio of the defective area, which is used to reflect the impact of the defect on the integrity of the coating; the boundary pixels of the defective area are extracted, and the changes in boundary details are analyzed, including the frequency of detail changes and the number of mutation points, to quantify the complexity of the boundary. The boundary complexity coefficient is calculated by the ratio of the detail change frequency and the number of mutation points to the total length or number of points of the boundary to comprehensively evaluate the complexity of the boundary; based on the area ratio and perimeter information, the shape regularity deviation value of the defect area is calculated to quantify the regularity of the defect shape. The defect is assumed to be an ideal circle through the area, and its theoretical minimum perimeter is calculated and compared with the actual perimeter. The larger the deviation value, the more irregular the defect shape. At the same time, the detail change frequency and the number of mutation points of the defect boundary are combined to further quantify the complexity of the boundary. The detail change frequency reflects the continuous change of the boundary line, and the number of mutation points captures the points of sharp change on the boundary, which are comprehensively used to calculate the boundary complexity coefficient; through normalization and weighted processing, the shape regularity deviation value and the boundary complexity coefficient are combined to generate a geometric influence factor. Normalization ensures that the two are consistent in dimension. The weighted combination adjusts the influence ratio of regularity and complexity according to actual detection needs, and finally generates a key indicator reflecting the influence of defect geometric characteristics on coating performance. The geometric influence factor comprehensively quantifies the defect shape and boundary characteristics.

[0123] Embodiment 2;

[0124] Based on the same inventive concept as the PVC floor surface coating defect detection method in the aforementioned embodiment, the present invention also provides a PVC floor surface coating defect detection system, which includes:

[0125] The area scanning module divides the coating surface to be tested into areas, obtains area scanning images, and performs image optical compensation on the area scanning images;

[0126] The defect selection module analyzes the high-dimensional image features of the scanned image of each area, performs difference analysis between the high-dimensional image features and the reference features, and selects the surface defect areas;

[0127] Type judgment module: establishes a high-dimensional defect feature analysis model, analyzes and generates defect location information and defect geometric parameters, and judges the defect type of the surface defect area based on the defect location information and defect geometric parameters;

[0128] The performance evaluation module quantitatively evaluates the sensitivity weight of the defect to the surface performance of the coating to be tested based on the defect location information and defect geometric parameters, and outputs the overall surface performance impact analysis results.

[0129] The above-mentioned adjustment system in the present invention can effectively implement the method for detecting defects in the surface coating of PVC flooring, and the technical effects that can be achieved are as described in the above-mentioned embodiments and will not be repeated here.

[0130] More specifically, the defect selection module includes:

[0131] Feature decomposition unit, which performs multi-scale feature decomposition on the area scan image and extracts high-dimensional image features including texture features, edge features and brightness distribution features;

[0132] The potential marking unit compares the high-dimensional image features with the reference features to train and generate a difference analysis algorithm. Based on the difference analysis algorithm, the difference index is calculated and the potential defect area is identified.

[0133] Connectivity analysis unit, which performs regional connectivity analysis on potential defect areas based on benchmark features and generates preliminary surface defect areas;

[0134] The optimization and adjustment unit makes fine adjustments to the boundaries of the preliminary surface defect area based on the difference analysis algorithm to generate the surface defect area.

[0135] Similarly, the above-mentioned optimization schemes for the system can also respectively achieve the corresponding optimization effects of the method in Example 1, which will not be repeated here.

[0136] Although the present application has been described with reference to specific features and embodiments thereof, it is apparent that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and drawings are merely illustrative of the present application as defined herein and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, the present application is intended to include such modifications and variations as fall within the scope of the present application and its equivalents.

Claims

1. A method for detecting defects in the surface coating of PVC flooring, characterized in that: The method comprises: Dividing the surface of the coating to be tested into regions, acquiring regional scanning images, and performing image optical compensation on the regional scanning images; Analyzing high-dimensional image features of the scanned image of each region, performing difference analysis between the high-dimensional image features and the reference features, and selecting the surface defect region; Establishing a high-dimensional defect feature analysis model, analyzing and generating defect location information and defect geometric parameters, and determining the defect type of the surface defect area based on the defect location information and the defect geometric parameters; According to the defect location information and the defect geometric parameters, a quantitative evaluation is performed on the sensitivity weight of the defect to the surface performance of the coating to be tested, and an overall surface performance impact analysis result is output; Determining the defect type of the surface defect area includes: Generating the high-dimensional defect feature analysis model according to the high-dimensional image features and the reference feature training; Scanning the surface defect area pixel by pixel to generate a pixel scanning result; Calculate the center point coordinates and area boundary information of the defect position based on the pixel scanning result to generate the defect location information; Calculating the area, perimeter, and shape regularity of the defect based on the pixel scanning result and the defect location information to generate the defect geometric parameters; Associating the defect location information and the defect geometric parameters based on the high-dimensional defect feature analysis model to determine the defect type of the surface defect area; The sensitivity weight of the surface performance of the coating to be tested is quantitatively evaluated in combination with defects, including: Determining the relative position distribution of the surface defect area in the total area of ​​the coating surface to be tested according to the defect location information, and calculating the area position weight; Calculating a geometric influence factor based on the defect geometric parameters, the geometric influence factor including the influence of the area ratio on the overall integrity and the shape regularity deviation value; Calculating a sensitivity weight by combining the regional position weight and the geometric influence factor, wherein the sensitivity weight is used to quantify the potential impact of the defect on the coating performance; For each of the surface defect areas, calculating a performance impact value of the single surface defect area according to the sensitivity weight and the geometric impact factor; Cumulatively calculating the performance impact values ​​of all the surface defect areas to generate an overall surface performance impact analysis result; Calculate the geometric impact factor, including: Obtaining the area and perimeter of the surface defect region, and calculating the area ratio and boundary complexity of the defect region; Calculating a shape regularity deviation value of the defective area based on the area ratio of the defective area, wherein the shape regularity deviation value is determined based on a ratio of an actual perimeter of the defective area to a theoretical minimum perimeter; Calculating a boundary complexity coefficient according to the boundary complexity of the surface defect area and based on the detail change frequency and the number of mutation points of the defect boundary; The geometric impact factor is generated by combining the shape regularity deviation value and the boundary complexity coefficient.

2. The method for detecting defects in the surface coating of PVC flooring according to claim 1, characterized in that: Select surface defect areas, including: Performing multi-scale feature decomposition on the area scan image to extract the high-dimensional image features including texture features, edge features, and brightness distribution features; Comparing the high-dimensional image features with the reference features, training and generating a difference analysis algorithm, calculating difference indicators based on the difference analysis algorithm and identifying potential defect areas; Performing regional connectivity analysis on the potential defect region based on the reference features to generate a preliminary surface defect region; The boundary of the preliminary surface defect area is finely adjusted based on the difference analysis algorithm to generate the surface defect area.

3. The method for detecting defects in the surface coating of PVC flooring according to claim 2, characterized in that: Performing a regional connectivity analysis on the potential defect region based on the benchmark features includes: Performing neighborhood connectivity calculation on the pixels within the potential defect area based on pixel grayscale values ​​and geometric structure continuity to screen out defective pixel groups that meet the connectivity requirements; Performing cluster analysis on the defective pixel group and further grouping them to generate multiple connected sub-regions and calculating texture consistency of the connected sub-regions; Performing connectivity analysis on the morphological parameters of the connected sub-regions, and eliminating pseudo-defect regions where the texture consistency is lower than a set value based on the reference feature comparison; According to the results of the connectivity analysis, a preliminary surface defect region is generated.

4. The method for detecting defects in the surface coating of PVC flooring according to claim 3, characterized in that: Calculating the texture consistency of the connected sub-regions includes: Analyzing the connected sub-regions and obtaining texture direction features and texture frequency features of the connected sub-regions; Calculating and quantifying the consistency of the texture direction within the connected sub-region based on the texture direction feature to generate a direction standard deviation; Calculating the complexity of texture details of the connected sub-region based on the texture frequency feature to generate a texture spectrum distribution deviation; The direction standard deviation and the texture spectrum distribution deviation are weighted and comprehensively calculated to comprehensively generate the texture consistency of the connected sub-region.

5. The method for detecting defects in the surface coating of PVC flooring according to claim 1, characterized in that: Calculating the center point coordinates and area boundary information of the defect position based on the pixel scanning result includes: Performing pixel classification screening on the pixel scanning results, marking pixels that meet a preset difference threshold as defective pixels, and generating a defective pixel set; Performing connectivity analysis on the defective pixel set to screen out defective areas that meet connectivity rules; Calculating a minimum circumscribed rectangle of the surface defect area according to pixel distribution of the surface defect area; Calculating the center point coordinates of the defect position based on the geometric center of the minimum circumscribed rectangle; Boundary information of the minimum circumscribed rectangle is obtained to generate region boundary information of the surface defect region. 6.PVC floor surface coating defect detection system, characterized by: The method for detecting defects in the surface coating of a PVC floor according to claim 1 is used, wherein the system comprises: The area scanning module divides the coating surface to be tested into areas, obtains area scanning images, and performs image optical compensation on the area scanning images; The defect selection module analyzes the high-dimensional image features of the scanned image of each area, performs difference analysis between the high-dimensional image features and the reference features, and selects the surface defect areas; Type judgment module: establishes a high-dimensional defect feature analysis model, analyzes and generates defect location information and defect geometric parameters, and judges the defect type of the surface defect area based on the defect location information and defect geometric parameters; The performance evaluation module quantitatively evaluates the sensitivity weight of the defect to the surface performance of the coating to be tested based on the defect location information and defect geometric parameters, and outputs the overall surface performance impact analysis results.

7. The PVC floor surface coating defect detection system according to claim 6, characterized in that: The defect selection module includes: Feature decomposition unit, which performs multi-scale feature decomposition on the area scan image and extracts high-dimensional image features including texture features, edge features and brightness distribution features; The potential marking unit compares the high-dimensional image features with the reference features to train and generate a difference analysis algorithm. Based on the difference analysis algorithm, the difference index is calculated and the potential defect area is identified. Connectivity analysis unit, which performs regional connectivity analysis on potential defect areas based on benchmark features and generates preliminary surface defect areas; The optimization and adjustment unit makes fine adjustments to the boundaries of the preliminary surface defect area based on the difference analysis algorithm to generate the surface defect area.

Images