Training method and system applied to printing defect detection model

Abstract

The invention relates to the technical field of printing defect detection, in particular to a training method and system applied to a printing defect detection model. The method comprises the following steps: building a printing defect detection model by collecting appearance data of a printed finished product and performing image optimization processing; analyzing a distortion abnormal state and a detail gradient fuzzy condition of the printed pattern through the optimized image data, and evaluating a visual performance decline situation; determining the abnormal running condition of the printing plate, positioning the defects of the printing plate, and evaluating the structural aging degree of the printing plate; predicting the error growth trend of the printing plate according to the aging degree, and recognizing printing abnormal defects in combination with pattern visual performance decline; carrying out optimization training on a defect detection model by utilizing the defect conditions; according to the invention, the printing defect detection model is trained, so that the printing defect detection is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of printing defect detection, and in particular to a training method and system for a printing defect detection model. Background Art

[0002] Print quality inspection methods typically rely on manual visual inspection, which struggles to fully identify complex and diverse printing defects. This is particularly true for the continuous inspection of high-precision, complex patterns, where significant limitations exist. Defect types in printed products include blurred patterns, scale distortion, missing details, structural distortion, image and text stretching and compression, and color block misalignment. These defects affect the aesthetics and recognizability of the finished product, leading to subsequent issues such as loss of traceability information and automated system recognition failures. Computer vision and deep learning algorithms are being gradually introduced to improve the efficiency and accuracy of printing defect detection and to build defect detection models suitable for different scenarios. Existing inspection systems lack dynamic update and self-optimization mechanisms, resulting in model training remaining static and unable to adapt to the gradual evolution of factors such as plate wear and stress fatigue caused by long-term changes in printing conditions. However, traditional printing defect detection suffers from inaccurate assessments of the decreasing visual performance of printed patterns and inaccurate detection of abnormal printing defects. Summary of the Invention

[0003] Based on this, it is necessary to provide a training method and system for a printing defect detection model to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a training method for a printing defect detection model includes the following steps:

[0005] Step S1: Acquire appearance data of a printed product; perform image capture on the appearance data of the printed product to obtain image data of the printed product; perform image optimization processing on the image data of the printed product to obtain optimized image data of the printed product; construct a printing defect detection model based on the optimized image data of the printed product to obtain a printing defect detection model;

[0006] Step S2: determining the abnormal distortion state of the printed pattern based on the optimized data of the printed product image; detecting the gradient blurring state of the printed pattern details based on the optimized data of the printed product image; and evaluating the decreasing trend of the visual performance of the printed pattern based on the gradient blurring state of the printed pattern details and the abnormal distortion state of the printed pattern;

[0007] Step S3: determining an abnormal operation condition of the printing plate based on the decreasing trend of the visual performance of the printed pattern; performing printing plate defect location processing based on the abnormal operation condition of the printing plate to obtain printing plate defect location data; and evaluating the degree of structural aging of the printing plate based on the printing plate defect location data;

[0008] Step S4: predicting the growth trend of the progressive error of the printing plate according to the aging degree of the printing plate structure; determining the abnormal printing defect situation according to the growth trend of the progressive error of the printing plate and the decreasing trend of the visual performance of the printed pattern; using the abnormal printing defect situation to optimize the model training of the printing defect detection model to obtain the printing defect detection optimization training model.

[0009] This invention effectively improves the accuracy of printing defect detection by accurately collecting and optimizing data on the appearance of finished printed products. Image optimization reduces image noise caused by external environmental or equipment factors, ensuring the accuracy of the defect detection model. Furthermore, detecting abnormal distortion of printed patterns and blurred gradients of details provides a basis for assessing the deterioration of the visual performance of printed patterns, thereby revealing potential issues and development trends in printing quality. By detecting abnormal operating conditions of printing plates and assessing the degree of structural aging, potential faults in printing plates can be promptly identified and defects accurately located, preventing the release of defective products. Structural aging analysis provides a predictive estimate of performance degradation over long-term use, facilitating the development of maintenance and replacement plans. Further incorporating the prediction of progressive error growth trends, precise monitoring and early warning of printing plates is achieved. By accurately identifying and optimizing defect detection models, the stability and accuracy of the printing production line are enhanced. This system not only optimizes existing inspection processes but also significantly improves the efficiency and quality of printing defect detection, ensuring consistent product quality and maximizing production efficiency. Therefore, the present invention is an optimization process for the traditional training of printing defect detection models, which solves the problems of inaccurate assessment of the decreasing trend of visual performance of printed patterns and inaccurate detection of abnormal printing defects in the traditional training of printing defect detection models, and improves the accuracy of assessment of the decreasing trend of visual performance of printed patterns and the accuracy of detection of abnormal printing defects.

[0010] The present invention also provides a training system for a printing defect detection model, which is used to execute the training method for a printing defect detection model as described above. The training system for a printing defect detection model includes:

[0011] A model building module is used to obtain appearance data of a printed product; perform image acquisition of the printed product on the appearance data of the printed product to obtain image data of the printed product; perform image optimization processing on the image data of the printed product to obtain optimized image data of the printed product; and construct a printing defect detection model based on the optimized image data of the printed product to obtain a printing defect detection model;

[0012] A visual performance degradation trend assessment module is used to determine the abnormal distortion state of the printed pattern based on the optimized data of the printed product image; detect the gradient blurring state of the printed pattern details based on the optimized data of the printed product image; and assess the visual performance degradation trend of the printed pattern based on the gradient blurring state of the printed pattern details and the abnormal distortion state of the printed pattern;

[0013] A structural aging assessment module is used to determine abnormal operation conditions of the printing plate based on the decreasing trend of visual performance of the printed pattern; to perform printing plate defect location processing based on the abnormal operation conditions of the printing plate to obtain printing plate defect location data; and to assess the structural aging degree of the printing plate based on the printing plate defect location data;

[0014] The model training optimization module is used to predict the growth trend of the progressive error of the printing plate according to the aging degree of the printing plate structure; determine the abnormal printing defects according to the growth trend of the progressive error of the printing plate and the decreasing trend of the visual performance of the printed pattern; use the abnormal printing defects to optimize the model training of the printing defect detection model to obtain the printing defect detection optimization training model.

[0015] The present invention is applied to a training system for a printing defect detection model. The system can implement any training method of the present invention applied to a printing defect detection model, and is used to combine the operation and signal transmission medium between each module to complete the training method applied to the printing defect detection model. The internal modules of the system cooperate with each other, and through the precise collection and optimization of printed product image data, combined with visual performance degradation, printing plate aging evaluation and progressive error prediction, the accuracy and efficiency of printing defect detection are effectively improved, the defect detection model is optimized, and the printing quality control capability is enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic flow chart of the steps of a training method for a printing defect detection model;

[0017] Figure 2 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0018] Figure 3 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0019] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0020] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0021] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0022] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0023] To achieve this, please refer to Figures 1 to 3 A training method for a printing defect detection model includes the following steps:

[0024] Step S1: Acquire appearance data of a printed product; perform image capture on the appearance data of the printed product to obtain image data of the printed product; perform image optimization processing on the image data of the printed product to obtain optimized image data of the printed product; construct a printing defect detection model based on the optimized image data of the printed product to obtain a printing defect detection model;

[0025] In an embodiment of the present invention, an industrial-grade linear array CCD high-speed camera and a backlight module are deployed to perform synchronous image acquisition in a fixed detection section of the conveyor belt path of the printed product. The PLC control module is used to synchronously control the image acquisition rhythm and the conveyor belt speed to ensure that each printed product is in a state of no displacement or shaking during acquisition. The image acquisition format is set to 16-bit grayscale image and three-channel RGB image parallel acquisition, and the image resolution is set to 0.01mm / pixel to ensure that subsequent image features can be accurately separated. The collected image data is transmitted to the edge computing node via Gigabit Ethernet. The image optimization processing process relies on OpenCV and Halcon image processing libraries. A high-pass filter is used in the edge node to sharpen the printed image, and a multi-scale Retinex algorithm is used to dynamically adjust the image brightness unevenness and color saturation imbalance problems. After image enhancement, Canny edge extraction and morphological opening operation algorithms are used to remove image noise to generate "printed product image optimization data" with clear edges and balanced colors, and stored in .tiff lossless format. The printing defect detection model is constructed. Based on the geometric distribution characteristics of the registration mark area, printing boundary area, and pattern entity area contained in the image optimization data, a regional positioning module for defect detection is constructed using image geometric matching and contour recognition methods. This module implements pre-screening logic for abnormal areas such as uneven color block area, detached graphic boundaries, and localized line breaks by setting specific threshold rules. Using the regional characteristics as the input source, a defect type judgment process is constructed and a "printing defect detection model" is generated. The model structure is established based on a rule logic tree and does not involve any AI learning mechanism. The model output includes: abnormal area coordinates, abnormal type classification labels, and boundary offset data, all of which are saved as structured data files for subsequent processing.

[0026] Step S2: determining the abnormal distortion state of the printed pattern based on the optimized data of the printed product image; detecting the gradient blurring state of the printed pattern details based on the optimized data of the printed product image; and evaluating the decreasing trend of the visual performance of the printed pattern based on the gradient blurring state of the printed pattern details and the abnormal distortion state of the printed pattern;

[0027] In an embodiment of the present invention, the pixel arrangement difference between the main visual area and the edge reference area of the printed pattern is extracted from the "printed product image optimization data", and the spatial offset between the actual printed pattern and the design reference pattern is analyzed by pattern template matching and image coordinate alignment methods. In the specific implementation process, the template transformation function in the Halcon image matching tool is used to perform affine matching processing on the original template, and then combined with corner point recognition and translation vector calculation to generate "printed pattern distortion abnormality state" data. This data quantifies the area, offset direction, offset (in mm) and offset slope gradient (in degrees) where the current pattern undergoes geometric deformation. Then, the detail gradient blurring condition is analyzed, and the image high-frequency gradient energy detection method based on Laplace transform is used to extract the texture clarity parameters of the key details of the pattern (such as border curves, text strokes, and color overprinting areas). If the gradient energy value is lower than the set threshold (the threshold is determined to be below 14.2 unit intensity based on experimental statistics), it is determined to be "printed pattern detail gradient blurring". This type of gradient intensity analysis is performed on multiple detail areas in the image and a blur condition identification map is generated, recording the coordinates, area, gradient distribution difference and relative standard deviation of each blur area. Combining the above two data, namely the abnormal state of printed pattern distortion and the blur condition of details, the visual performance decline trend of the printed pattern is evaluated according to the level classification standard of visual performance attenuation. The specific method includes constructing an abnormal area ratio index (the proportion of defective areas per unit area), a detail loss rate index (detail area gradient attenuation ratio) and a color gamut deviation index (color block area mean deviation) to form a complete "printed pattern visual performance decline trend" evaluation data set, which provides a basis for the next step of judging the abnormal state of the printing plate.

[0028] Step S3: determining an abnormal operation condition of the printing plate based on the decreasing trend of the visual performance of the printed pattern; performing printing plate defect location processing based on the abnormal operation condition of the printing plate to obtain printing plate defect location data; and evaluating the degree of structural aging of the printing plate based on the printing plate defect location data;

[0029] In an embodiment of the present invention, based on the "printing pattern visual performance decrease trend" evaluation data generated in the previous step, an inference and judgment operation of the abnormal operation state of the printing plate is performed. This operation is based on the pattern attenuation characteristic parameter comparison table corresponding to different plate wear degrees in historical printing batches, and a parameter lookup table is deployed in the edge node for dynamic matching. By setting the threshold interval, the current pattern visual attenuation trend data is mapped to the plate operation status level, such as "mild abnormality", "moderate wear", "severe offset", etc., to generate "printing plate operation abnormality" marking data. According to the abnormal operation status information, combined with the printing pattern area coordinates and graphic projection errors recorded in the printed product image optimization data, the printing plate defect positioning processing is performed. Using a mapping algorithm based on geometric inversion, the pattern deformation area is reversely projected into the printing roller plate coordinate system, and the potential damage source is located through contour mapping identification. The center coordinates of the defect area, coverage range, error gradient direction, etc. form the "printing plate defect positioning data" and are marked on the plate expansion diagram to assist in subsequent maintenance. Based on the distribution density of each defect location in the defect location data and the historical cumulative area ratio, the trend of micro-wear of the plate surface material, the tendency of pressure-resistant coating shedding, and the direction of plate deformation are analyzed to calculate the "degree of aging of the printing plate structure." This degree is expressed by an aging index, which is weighted by the cumulative damage area, the number of damages, and the consistency coefficient of the damage direction. The index ranges from 0 to 1, with higher values indicating more severe aging. This result will be combined with subsequent error trend assessment.

[0030] Step S4: predicting the growth trend of the progressive error of the printing plate according to the aging degree of the printing plate structure; determining the abnormal printing defect situation according to the growth trend of the progressive error of the printing plate and the decreasing trend of the visual performance of the printed pattern; using the abnormal printing defect situation to optimize the model training of the printing defect detection model to obtain the printing defect detection optimization training model.

[0031] In an embodiment of the present invention, the parameter "degree of aging of the printing plate structure" is read, and the stage of progressive error growth of the current printing plate is determined based on its value range. The feedforward differential incremental analysis method is used to calculate the incremental change trend of the pattern position error, pattern repeatability error (repeatability) and pattern stretch deformation error during the continuous printing process. With the time axis as the horizontal axis and the error amount as the vertical axis, the error growth trend curve is drawn, and the trend slope and acceleration are extracted to obtain the indicator data of "progressive error growth trend of the printing plate". The trend data is cross-compared with the previously obtained "declining trend of visual performance of the printed pattern" to construct a comprehensive mapping table of abnormal defects containing time series and spatial superposition. Combining the error source location and the degree of visual abnormality, accurate judgment of "abnormal printing defect conditions" is achieved. Such defect conditions include: periodic pattern stretching, increased lateral boundary blur, double pattern ghosting, and continuous increase in boundary line defects. The above-mentioned printing abnormal defect condition data is fed back to the initially constructed "printing defect detection model" to achieve model training optimization processing. Without relying on deep learning, the original model's region recognition and defect classification modules were restructured using three mechanisms: defect frequency weight adjustment, dynamic adaptive threshold adjustment, and enhanced region discrimination logic. New recognition features, anomaly ranges, structural aging patterns, and error trend data were incorporated into the judgment rule table as conditional logic, generating an "optimized training model for printing defect detection." This optimized model was output in JSON format and loaded into the system's operating environment as an updated configuration for the real-time printing defect monitoring system.

[0032] Preferably, step S1 includes the following steps:

[0033] Step S11: setting the resolution of the visible light camera to 1920×1080 pixels, the minimum light sensitivity to 0.01 lux, and the frame rate to 30 FPS;

[0034] In an embodiment of the present invention, in the image acquisition link of the printed product, in order to ensure that the image detail information is completely captured, an industrial-grade visible light camera with a model of SONY IMX327 sensor is selected, and its image acquisition resolution is set to 1920×1080 pixels to ensure that the image details have a clarity of at least 300 pixels per square millimeter. In order to improve the imaging capability in low-light environments, the camera's minimum light sensitivity is set to 0.01lux to ensure that high-quality image acquisition can still be completed under insufficient lighting conditions in the workshop (such as less than 30lux). At the same time, the frame rate is set to 30 frames per second (30FPS) to avoid image tearing or acquisition delays while maintaining image continuity and dynamic response performance. The camera is connected to the main control processing system via a USB 3.0 industrial camera interface, and uses the DirectShow protocol for data stream acquisition to ensure acquisition stability. This configuration is intended to provide hardware support for subsequent image data acquisition and image processing links to ensure that data quality meets standardized processing requirements.

[0035] Step S12: Obtaining appearance data of the printed product;

[0036] In an embodiment of the present invention, the acquisition of the appearance data of the finished printed product relies on the visible light camera acquisition system with the above configuration. A collection window is set on the production line, and an image collection station is set at the first inspection position after printing is completed. Each time the finished product passes this position, the system triggers a shooting instruction to record its front image. The area of the acquisition area is set to A4 paper size (210mm×297mm), and the image is calibrated by an optical calibration device to eliminate the edge deformation error caused by lens distortion. During the acquisition process, an angle adjustment platform is additionally provided so that the sample image acquisition angle is controlled within ±2° vertically to avoid the impact of angle offset on the accuracy of subsequent image processing. The acquired appearance data of the printed product is stored in the original RGB format. The pixel array size corresponding to each image is 1920×1080, and the color bit depth is 8bit. It is uploaded to the image processing server through the frame buffer mechanism.

[0037] Step S13: collecting the printed product image data based on the appearance data of the printed product using the visible light camera to obtain the printed product image data;

[0038] In an embodiment of the present invention, after obtaining the basic data of the appearance image, the image acquisition and processing flow is entered. During the execution of the acquisition link, the image acquisition control unit is used to perform signal scheduling on the camera, and the production rhythm is synchronously controlled in combination with the PLC (programmable logic controller) to ensure that the image acquisition accurately matches the time when the finished product passes the inspection point. The acquired image undergoes preliminary mean filtering to remove impulse noise, and then the image is subjected to ROI (Region of Interest) area extraction operation. According to the set image coordinate parameters, non-printing areas such as borders, binding holes and other interference information are stripped from the image, and only the core printing pattern area is retained. The extracted image is stored in lossless PNG format through the image compression function of the OpenCV image processing library, and is accompanied by metadata such as timestamp, product batch number, printing task number, etc., to construct a printed product image dataset as the input source for subsequent image optimization processing.

[0039] Step S14: performing image optimization processing on the printed product image data to obtain the printed product image optimized data;

[0040] In an embodiment of the present invention, during the image optimization process, the collected printed product image data is subjected to brightness normalization processing to eliminate the pixel grayscale offset caused by changes in illumination. The histogram equalization algorithm (HistogramEqualization) is used to adjust the overall brightness of the image to enhance the contrast between pattern boundaries and layers. Subsequently, the Sobel edge enhancement processing method is used to sharpen the image contour, retain the key printed pattern feature edges and remove background noise interference. For artifact noise and grayscale blurred areas in the image, a bilateral filter algorithm (BilateralFilter) is used to perform edge-preserving filtering processing to avoid blurring and erasing of detailed structures. After completing the above operations, the structural similarity detection algorithm (SSIM) is used to compare with the standard image template to verify whether the clarity of the optimized image reaches the set threshold (for example, SSIM>0.85) and output the printed product image optimization data, and store it in a standardized format with a uniform image size and image bit depth for use in defect detection tasks.

[0041] Step S15: constructing a printing defect detection model based on the optimized data of the printed product image to obtain a printing defect detection model.

[0042] In this embodiment of the present invention, based on the optimized image data of the finished printed product generated in step S14, the image feature set required for the print defect detection model is constructed. A grayscale gradient distribution analysis method is used to extract the edge distribution map within each image. An image texture feature matrix is then constructed by combining statistical features such as the grayscale distribution mean and standard deviation. Subsequently, a 10×10 pixel sliding window analysis technique is used to scan the entire image area. The degree of color shift within the local block, boundary fracture characteristics, and local texture discontinuities are statistically analyzed to form a corresponding defect label coordinate set. These feature data and image coordinates are annotated one by one and organized into a structured training sample set in matrix form. A feature rule set-based discriminant method is used to construct the model. Standard defect types defined by the printing industry (such as "broken line," "burr," "overprint offset," "ghosting," and "color spot") are used as output labels, and the degree of feature anomaly in the corresponding image area is used as an input judgment factor to construct a defect detection model in the form of a decision tree. After the model is constructed, the model structure and parameter rule set are saved in JSON format to serve as the basis for subsequent training optimization and anomaly detection processes.

[0043] Preferably, step S14 includes the following steps:

[0044] Step S141: performing a gradient amplitude distribution analysis on the printed product image data to obtain a gradient amplitude distribution of the printed product image;

[0045] In an embodiment of the present invention, gradient amplitude analysis is performed on the image data of the finished printed product, and the calculation is performed using the Sobel operator in the image processing software package OpenCV. The Sobel operator calculates the gradient amplitude of each pixel by filtering the image in the horizontal and vertical directions. The gradient amplitude reflects the intensity of the edges and textures in the image and can effectively identify image details and contours. In this process, the original image is grayscaled to remove color information and retain only brightness information to reduce computational complexity. Then, the image is convolved using the Sobel operator to calculate the gradient amplitude of each pixel in the image to obtain the gradient information of each pixel. The calculation results will be stored in the form of a two-dimensional matrix, where the values of the matrix represent the gradient amplitude of each pixel in the image. The gradient amplitude data obtained from this analysis will serve as the basis for subsequent image clarity evaluation.

[0046] Step S142: evaluating the clarity of the printed product image based on the gradient amplitude distribution of the printed product image being greater than 60;

[0047] In an embodiment of the present invention, based on the gradient amplitude distribution of the printed product image obtained in step S141, the image clarity is further evaluated, and the average and standard deviation of the gradient amplitudes of all image pixels are calculated. For the gradient amplitudes of all pixels, their distribution is statistically analyzed to obtain the overall "sharpness" or "clarity" of the image. When the average gradient amplitude of the image is greater than 60, it is considered that the clarity of the image is at a high level. In the specific implementation process, Python scripts are used for automated calculations, and the gradient amplitudes of all pixels are summed and the standard deviation is calculated using the Numpy library. If the result meets the conditions (i.e., the gradient amplitude is greater than 60), the image is evaluated as clear; if not, it is evaluated as unclear. The clarity evaluation value serves as one of the indicators for subsequent image quality detection.

[0048] Step S143: performing local texture intensity recognition of the printed product image based on the printed product image data to obtain local texture intensity data of the image;

[0049] In an embodiment of the present invention, the identification of the local texture intensity of an image is performed using image segmentation and local feature extraction technology. The image is divided into multiple small blocks, and a block size of 8×8 pixels is usually used for local analysis. For each image block, a local binary pattern (LBP) algorithm is used for feature extraction. The LBP algorithm generates a binary pattern to describe the texture characteristics of the area by calculating the relationship between the values of the neighboring pixels around each pixel point. The same operation is performed on all image blocks to obtain the texture intensity data of each image block. Then, through statistical analysis of these local texture intensity data (such as mean and standard deviation), the local texture intensity characteristics of the image are further obtained. These data will help determine whether there is texture damage or loss in the local area of the image, thereby providing a basis for image quality assessment.

[0050] Step S144: evaluating the initial quality of the printed product image based on the image local texture intensity data and the clarity of the printed product image;

[0051] In an embodiment of the present invention, based on the results of steps S142 and S143, the initial quality of the printed product image is further evaluated. In the specific implementation process, the comprehensive quality score of the image is calculated by combining the analysis of image clarity (the result of step S142) and local texture intensity (the result of step S143). The quality score is calculated using a weighted average method, in which clarity accounts for 70% and texture intensity accounts for 30%. For example, for an image with a clarity score greater than 60, if its local texture intensity is within the specified range, the image quality is considered to be good; if the local texture intensity is too low, there are printing defects or quality problems, and the image needs to be further optimized. The specific scoring process is implemented by writing a Python script, which synthesizes the clarity score and texture intensity score into a quality score according to the set weights, and outputs a comprehensive quality score value, which is used to determine whether the printed product image meets the preset quality standards.

[0052] Step S145: Dividing the initial quality of the printed product image into 2200 grid regions to obtain grid region data of the printed product image quality;

[0053] In an embodiment of the present invention, in this step, the initial quality of the printed product image is gridded and divided into regions. The image is divided into 2200 uniform regions, each of which is approximately 4.8×4.8 mm in size, which is suitable for standard images of A4 paper. The division process sets a fixed-size grid template and scans the entire image in sequence for cutting, ensuring that each region is the same size and covers every part of the image. Through this operation, the quality information of the image is refined into each region. After the division is completed, for each grid region, a quality score is assigned to each region based on the previous clarity and texture intensity evaluation results. In this way, the obtained grid region data provides detailed local quality information for the subsequent detection of high-frequency component loss areas.

[0054] Step S146: detecting the high-frequency component loss area of the finished product image based on the quality grid area data of the printed product image;

[0055] In an embodiment of the present invention, the quality grid area data of the printed product image obtained in step S145 is used to detect the areas where high-frequency components are lost in the image, and the frequency domain analysis is performed on each grid area using Fourier transform to convert it into an image in the frequency domain. The Fourier transform converts the spatial information of the image into frequency information, which facilitates the analysis of the high-frequency components of the image (such as details, textures, etc.). For each area, the energy distribution of the high-frequency part of its spectrum is calculated. If the high-frequency components of certain areas are obviously lost, it indicates that the area is blurred, noisy or has printing defects. In this process, the Fast Fourier Transform (FFT) algorithm is used to quickly complete the frequency domain conversion, and Python is used to implement automated frequency analysis. The criterion for determining an area where high-frequency components are lost is that the frequency energy of the area is less than a set threshold.

[0056] Step S147: performing image optimization processing on the high-frequency component missing area of the finished product image to obtain optimized data of the printed finished product image.

[0057] In an embodiment of the present invention, image optimization processing is performed on the area where high-frequency components are lost, and a high-pass filtering method is used to restore the lost image details. Specifically, for each area where high-frequency components are lost, local image enhancement processing is performed, and a Laplacian filter is applied to enhance the details of the image. The Laplacian filter is a second-order derivative operator that can highlight the edges and high-frequency information of the image. By adjusting the parameters of the filter, precise enhancement of the high-frequency part can be achieved. The enhanced image area is merged with part of the original image to generate an optimized image. This process is implemented through the filtering function in OpenCV, and ensures that the optimized image can restore high-frequency details without generating artifacts. The optimized image is used as the optimized data for the printed product image, ready to enter the subsequent defect detection and model training process.

[0058] Preferably, determining the abnormal state of printed pattern distortion in step S2 includes:

[0059] Optimize data collection based on the printed product image to obtain the external contour of the printed product pattern;

[0060] In an embodiment of the present invention, optimized printed product image data is obtained. The image is digitally captured using a high-resolution camera or scanner to ensure that the image has sufficient details and resolution. After acquiring the image, an edge detection algorithm (such as Canny edge detection) is used to process the image. The edge detection algorithm can highlight the contours of objects in the image and identify the edges in the image by analyzing the grayscale changes of the image. The Canny function in the OpenCV library is used to perform grayscale conversion on the image, and then the edges of the image are extracted through Canny edge detection. The outer contour in the image is the edge information obtained by the algorithm. The key operation of this step is to efficiently identify the outer contour data of the printed product pattern and prepare these data for the next step of proportional analysis.

[0061] Measure the overall proportion change of the printed product pattern based on the printed product image optimization data;

[0062] In an embodiment of the present invention, a geometric measurement method is used to evaluate the overall proportional change of the printed finished pattern, and the obtained contour data is converted into a geometric feature representation. The total area of the pattern and its aspect ratio (the ratio between length and width) are calculated through the contour data. According to the geometric dimension information of the image, the overall shape change of the printed pattern is measured. The aspect ratio of the pattern is calculated using the bounding box in the image. The aspect ratio formula is the ratio of the length of the long side to the length of the short side. By comparing the proportions of the original design drawing, the proportion change data of the pattern is obtained by calculating the difference in the aspect ratio between the current pattern and the design drawing. This data will serve as the basis for subsequent distortion detection, reflecting whether the pattern is deformed by stretching or compression.

[0063] Integrate the external contour of the printed product pattern and the overall proportion change of the printed product pattern to obtain the characteristic data of the printed product pattern;

[0064] In an embodiment of the present invention, the obtained external contour data and the obtained proportional change data are integrated, and the external contour data and the proportional change data are merged into a set of feature vectors. During the merging process, the external contour conditions include the shape characteristics of the contour (such as area, perimeter, edge smoothness, etc.), and the proportional change conditions reflect the offset of the aspect ratio. By aggregating these data into feature data in a unified format, the next step of analysis and processing can be carried out more conveniently. These integrated data are used for further identification of pattern features and subsequent image registration processes.

[0065] Acquire printed image structure design data, and perform image registration processing on printed product pattern feature data based on the printed image structure design data to obtain printed product image registration data;

[0066] In an embodiment of the present invention, the printed image structure data at the time of design is obtained. These data usually come from the design drawings or CAD files of the printing press, and record information such as the ideal size, position and proportion of the pattern. After obtaining these design data, the integrated printed product pattern feature data is processed based on image registration technology. Image registration uses an algorithm based on corner detection, such as SIFT (Scale Invariant Feature Transform) or ORB (Oriented FAST and Rotated BRIEF) algorithm, to extract feature points in the pattern that match the design data, and then match these feature points with corresponding points in the design image through the registration algorithm. During the image registration process, the printed image is adjusted using affine transformation or perspective transformation to ensure that it is aligned with the design image. The obtained printed product image registration data eliminates the deviation of the pattern caused by the shooting angle or deformation, making it more consistent with the design image, and providing accurate data for subsequent node offset detection.

[0067] Measure the node offset of the printed pattern based on the image registration data of the printed product;

[0068] In an embodiment of the present invention, the obtained registration data is used to measure the node offset. Pattern nodes generally refer to feature points at key positions in an image, such as corner points or geometric center points of a pattern. Image analysis software (such as OpenCV or Matlab) is used to perform node detection on the registered image to locate important nodes in the pattern. Then, the deviation between the actual node position of the pattern and the node position in the design image is calculated. The deviation is quantified by calculating the Euclidean distance of the node position, and a node offset data set is generated. If the value of the node offset is greater than a set threshold, it indicates that the pattern has a large deformation. Through this step, the local offset of the pattern can be quantified, and a basis for subsequent image deformation analysis can be provided.

[0069] Evaluate the abnormal aspect ratio deviation of the image based on the node deviation of the printed pattern;

[0070] In an embodiment of the present invention, based on the obtained node offset data, the abnormal offset of the image aspect ratio is further evaluated. By comparing the aspect ratio of the printed product pattern with the design pattern, if the magnitude of the offset exceeds the predetermined tolerance range, it is considered that the pattern has an aspect ratio abnormality. In a specific implementation, the percentage deviation of the aspect ratio is calculated by comparing the aspect ratio in the design drawing with the actual aspect ratio of the pattern after alignment. If the deviation exceeds the set threshold, the image is considered to have a serious aspect ratio abnormality. This step helps to determine whether the pattern is affected by mechanical errors or design deviations during the production process, and then determine the degree of deformation of the image.

[0071] Detect the stretch-compression degree of the printed pattern based on the abnormal deviation of the image aspect ratio;

[0072] In an embodiment of the present invention, a stretching or compression test of the printed pattern is performed based on the abnormal aspect ratio deviation. The degree of stretching or compression of the pattern is quantified by calculating the deviation of the aspect ratio. If the aspect ratio deviation of the pattern is large, it indicates that the pattern has a stretching or compression deformation. This step compares the actual aspect ratio of the pattern with the aspect ratio of the design drawing and uses a proportional calculation formula to derive the degree of stretching or compression of the pattern. For example, if the actual aspect ratio is greater than the design aspect ratio, it means that the pattern is stretched; if the actual aspect ratio is less than the design aspect ratio, it means that the pattern is compressed. This test helps to accurately locate the type of pattern deformation and provides data support for the judgment of abnormal distortion.

[0073] The abnormal state of printed pattern distortion is determined based on the degree of stretching and compression of the printed pattern and the abnormal deviation of the image aspect ratio.

[0074] In an embodiment of the present invention, the abnormal distortion state of the pattern is comprehensively judged based on the obtained stretching and compression degrees, as well as the abnormal deviation of the aspect ratio. Specifically, if the stretching and compression degrees exceed the predetermined tolerance range, or the deviation of the aspect ratio exceeds the set threshold, the pattern is considered to be distorted. At this time, the appearance of the pattern will undergo obvious geometric changes, leading to printing quality problems. In actual operation, by writing judgment rules and combining the abnormal data of stretching, compression and aspect ratio, the system can automatically detect whether the pattern is distorted and output a corresponding abnormal distortion state report. This step provides image quality assessment results and provides a decision-making basis for subsequent defect repair or adjustment.

[0075] Preferably, the detection of the gradient blur condition of the printed pattern details in step S2 includes:

[0076] Collect the texture features of the printed product image based on the optimized data of the printed product image;

[0077] In an embodiment of the present invention, a high-resolution camera or scanner is used to obtain image data of the printed product to ensure that the quality of the image is optimized, and an image processing tool is used to pre-process the image to eliminate noise and blur. The pre-processing steps of image processing include color normalization, denoising, sharpening, etc. Then, the optimized image is processed using a texture analysis algorithm to extract the texture features of the printed product. In this embodiment, Gabor filters and wavelet transforms are used to extract local texture features of the image. Gabor filters can capture texture details in different directions in the image, while wavelet transforms help analyze the frequency components of the image to obtain local detail information of the image. Texture features include the change pattern of grayscale values in the local area of the image, texture complexity, directionality, etc. These texture feature data will constitute the basic data for image analysis and will be used for subsequent operations.

[0078] Identify the energy attenuation trend of local texture in the printed product image based on its texture features;

[0079] In an embodiment of the present invention, the extracted texture feature data is used to further analyze the attenuation trend of the local texture energy in the printed product image, and the image is divided into multiple sub-blocks (for example, 8x8 pixel blocks) according to the grayscale changes of the image texture features. The texture energy of each sub-block is calculated, specifically by calculating the grayscale variance or the sum of the squares of the gradient amplitude in the local area. The texture energy of each sub-block reflects the richness of the texture details of the area. If the texture energy of a local area is significantly lower than that of the adjacent area, it means that the texture details of the area are lost or attenuated. Furthermore, by comparing the texture energy values of each sub-block, the area with the attenuation trend is identified. The intensity and range of the attenuation trend indicate the severity of the loss of image details.

[0080] Evaluating the degree of detail loss in printed patterns based on the energy attenuation trend of local texture in images

[0081] In an embodiment of the present invention, after identifying areas of local texture energy attenuation, it is next necessary to evaluate the degree of pattern detail loss in these areas. By performing a detailed analysis of each attenuation area, a quantitative analysis is performed using the texture features of the image and the degree of energy attenuation. If the texture energy value of a local area is low and the details of the area are relatively complex, it can be determined that the degree of detail loss in the area is high. The degree of detail loss can be quantified by the relative difference in texture energy values. A threshold is set, and when the difference exceeds the threshold, it is marked as severe detail loss. In this way, the overall detail loss of the printed pattern is comprehensively evaluated to obtain the specific degree of detail loss in the pattern.

[0082] Detect the continuity interruption of the printed pattern based on the degree of loss of printed pattern details and the trend of local texture energy attenuation in the image;

[0083] In an embodiment of the present invention, after obtaining the degree of pattern detail loss, the continuity of the pattern is then detected by analyzing the relationship between the detail-missing area and the local texture energy attenuation trend of the image. The continuity of a printed pattern refers to the smooth transition and continuity of the pattern throughout the image. When the detail loss in a local area reaches a certain level, the pattern becomes discontinuous or broken. By analyzing the distribution of the detail-missing area and combining it with the attenuation trend of the local texture energy, it is possible to determine whether the pattern has been broken. Specifically, when the texture energy attenuation in certain areas is more obvious and the degree of detail loss in these areas is high, it can be considered that the continuity of the pattern has been interrupted.

[0084] Identify the image texture direction vector offset based on the texture features of the printed product image;

[0085] In an embodiment of the present invention, the texture features of the printed product image are extracted, and then the texture direction vector offset in the image is analyzed based on these texture features. Specifically, the gradient direction of each local area in the image is calculated to obtain the texture direction vector of the area. By analyzing the texture direction vector of the image, the area with directional offset is identified. When the texture direction in the image deviates significantly from the texture direction of the expected design, it means that the texture of the area has offset, which usually causes the visual effect of the pattern to deteriorate or become irregular. The calculation of the texture direction vector can be achieved through local gradient operation. For example, the Sobel operator is used to perform gradient calculation on the image to obtain the directional information of each pixel point. Further evaluation and processing are performed on each texture direction offset area.

[0086] Determine the weakening of pattern details in the printed product based on the image texture direction vector offset and the degree of loss of printed pattern details;

[0087] In an embodiment of the present invention, the offset of the image texture direction is identified, and then the detail weakening of the printed pattern is further analyzed in combination with the obtained degree of pattern detail loss. The detail weakening is evaluated by combining the degree of offset of the texture direction vector and the degree of pattern detail loss. If the texture direction offset is large and the detail loss is severe, it can be considered that the details of the pattern have been significantly weakened. At this time, by calculating the combined impact of detail loss and directional offset in each area, the degree of detail weakening of the printed pattern can be obtained. This evaluation process is carried out by weighted averaging or weighted scoring, so that the detail loss and texture offset are quantified into a comprehensive value for subsequent defect detection.

[0088] The gradient blurring of printed pattern details is detected based on the weakening of printed pattern details and the interruption of printed pattern continuity.

[0089] In an embodiment of the present invention, the detail weakening of the printed product pattern is obtained, and then the detail gradient blurring of the printed pattern is evaluated in combination with the detected pattern continuity interruption. Gradient blurring refers to the transition of the detail part of the printed pattern becoming blurred, which is usually caused by the weakening of the detail and the continuity interruption of the pattern. In this step, by comparing the detail weakening and the continuity interruption of each area, combined with the gradient calculation (such as the image gradient amplitude), it is determined whether the pattern has the detail gradient blurring phenomenon. If the pattern shows detail loss and continuity interruption in multiple areas, and the gradient information of these areas is weak, it means that the detail gradient blurring of the pattern is high. Through this process, the detail gradient blurring of the printed pattern can be accurately detected, further providing data support for the printing defect detection model.

[0090] Preferably, the evaluation of the decreasing trend of the visual performance of the printed pattern in step S2 includes:

[0091] Detect the structural distortion of the printed pattern based on the gradient blur of the printed pattern details and the abnormal distortion of the printed pattern;

[0092] In an embodiment of the present invention, the gradient information of the printed pattern is obtained by image gradient calculation, and the detail gradient blur condition is then determined. Detail gradient blur refers to the blurred edge transition of the detail part of the pattern, which is usually caused by texture loss, interruption of pattern continuity or reduction of image resolution. By calculating the gradient amplitude of each area of the image, the detail clarity of each area is obtained, and then it is determined whether the pattern has gradient blur. When the gradient amplitude of certain areas of the pattern is lower than the preset threshold, it means that the visual clarity of the area is reduced and the details appear blurred. In addition, an image distortion detection algorithm is used to identify abnormal pattern distortion. Distortion anomalies can be detected by calculating the geometric distortion of the image, uneven lighting, etc. For example, the Hough transform is used to detect whether the geometric shape of the pattern is distorted, or the image correction algorithm is used to compare the deviation between the expected shape and the actual shape of the image. By combining the gradient blur condition with the distortion anomaly, it is detected whether the printed pattern has structural distortion, and then the structural integrity of the pattern is evaluated.

[0093] Determine the increasing difficulty trend of printed pattern structure recognition based on the printed pattern structure distortion;

[0094] In an embodiment of the present invention, the structural distortion of the pattern is obtained through distortion anomaly detection and detail gradient blur detection. Next, the recognition difficulty of the printed pattern is evaluated based on the structural distortion condition. The structural distortion of a printed pattern usually means that the geometric form or detail information of the pattern is damaged, which will directly increase the difficulty of pattern recognition. In this embodiment, the changing trend of the difficulty of pattern recognition is evaluated by analyzing the scope and degree of structural distortion and combining the geometric features of the pattern (such as the smoothness of the lines, the symmetry of the pattern, etc.). Specifically, for pattern areas with a higher degree of distortion, the recognition difficulty will increase significantly. By calculating the area and boundary deformation of the structural distortion area, a structural recognition difficulty index is obtained, which reflects the recognition difficulty of the entire printed pattern. When the structural distortion area is relatively wide, or the distortion is severe, the structural recognition difficulty index will gradually increase, indicating that the recognition difficulty continues to increase.

[0095] Predict the sharp decline in the decoding rate of printed patterns based on the increasing trend of printed pattern structure recognition difficulty;

[0096] In an embodiment of the present invention, based on the calculated structural recognition difficulty index, the decoding rate of the printed pattern is then predicted. The decoding rate refers to the ability of the system to successfully parse the printed pattern, which is affected by factors such as pattern details, structural distortion, and pattern clarity. When the structural distortion of the pattern is severe, the decoding rate of the pattern will drop sharply. In this embodiment, based on the growth trend of the recognition difficulty index, historical data and statistical methods are used to predict the attenuation of the decoding rate. For example, a relationship model is established based on existing sample data to model the correlation between the degree of structural distortion of the pattern and the decoding rate, and the change in the decoding rate is inferred based on the degree of distortion of the pattern. In specific operations, regression analysis and other methods are used to establish a functional relationship between the degree of distortion and the decoding rate. Through this method, it is possible to predict how the decoding rate of the printed pattern will decay under different degrees of distortion and derive the corresponding attenuation trend.

[0097] Detecting printed pattern information reading failures based on the sharp decline in printed pattern decoding rates and the increasing difficulty in identifying printed pattern structures;

[0098] In an embodiment of the present invention, after predicting the trend of a sharp decline in the decoding rate, it is next necessary to evaluate the situation of information reading failure. The information reading of a printed pattern generally depends on the clarity of the pattern and the accuracy of structural recognition. When the decoding rate declines sharply, it indicates that the recognizability of the pattern is reduced, resulting in a failure to read the information. The detection of information reading failure is completed by comparing the error rate and the success rate during the decoding process. If the decoding error rate increases significantly and the success rate continues to decline, it is determined that the information reading has failed. In this process, the pattern is decoded with the help of an image processing tool, and the number of times the pattern is successfully decoded and the number of failures are counted. The occurrence of information reading failure is detected based on the growth trend of the number of failures. Specifically, an image recognition algorithm (such as OCR, barcode decoding algorithm) is used to decode the pattern multiple times, and the success or failure of each decoding is tracked. If the number of reading failures exceeds a certain threshold, it indicates that there is a serious recognition problem with the pattern.

[0099] Predict the breakage of the printed pattern traceability chain based on the failure of printed pattern information reading and the sharp decline of printed pattern decoding rate;

[0100] In an embodiment of the present invention, based on the information reading failures and the sharp attenuation of the decoding rate, the next step is to predict the breakage of the traceability chain of the printed pattern. The traceability chain refers to the process by which the information obtained by decoding the pattern can be traced back to the source of the pattern. If the pattern decoding fails or the information reading fails, it will directly lead to the breakage of the traceability chain. In this step, the probability of the traceability chain breakage is judged by the ratio of decoding successes and failures, combined with the pattern decoding rate attenuation trend. Specifically, if the pattern decoding rate is lower than a preset threshold and the number of information reading failures increases sharply, it means that the traceability chain of the pattern will be broken. The prediction process obtains a risk assessment of the traceability chain breakage by analyzing the historical data of decoding successes and failures, as well as the impact of pattern distortion on the traceability chain.

[0101] The decreasing trend of printed pattern visual performance is evaluated based on the breaks in the printed pattern traceability chain and the failure in reading printed pattern information.

[0102] In an embodiment of the present invention, a comprehensive assessment of the visual performance degradation of a printed pattern is conducted by combining predicted traceability chain breaks and information read failures. This visual performance degradation reflects the overall decline in visual quality of the pattern during actual use due to distortion, decoding issues, and other factors. In this embodiment, a comprehensive assessment model is constructed by combining the risk of traceability chain breaks and information read failures to output a visual performance degradation index for the printed pattern. This index is calculated based on the degree of information read failure, the trend of decoding rate decay, and the integrity of the traceability chain. When the visual performance degradation index exceeds a certain threshold, it indicates that the visual performance of the pattern has significantly declined and cannot meet the requirements of practical applications.

[0103] Preferably, step S3 includes the following steps:

[0104] Step S31: determining an abnormal operation condition of the printing plate based on the decreasing trend of the visual performance of the printed pattern;

[0105] In an embodiment of the present invention, the visual performance degradation of an assessed printed pattern is used as input to infer abnormalities in the printing plate's operation by analyzing the pattern's visual performance degradation. This degradation includes factors such as blurred detail gradients, abnormal distortion, and decode rate degradation, which reflect defects and anomalies in the printing process. Based on the data analysis results, an evaluation model is constructed that can infer whether a printing plate is operating abnormally based on the degree of pattern distortion and recognition difficulty. During operation, the visual performance data of the current pattern is compared with the normal pattern data to determine whether significant degradation or distortion has occurred. If the visual quality of the pattern significantly decreases and the corresponding decode rate or probability of information read failure increases significantly, this indicates that the printing plate is operating abnormally. Image processing software is used to compare the structure and texture of the pattern, calculate the degree of visual degradation in each area, and combine this with known pattern performance standards to determine whether the printing plate is operating abnormally. This process can determine the abnormal state of the printing plate, providing a basis for subsequent defect location and degradation prediction.

[0106] Step S32: performing printing plate defect location processing using the printing plate operation abnormality to obtain printing plate defect location data;

[0107] In an embodiment of the present invention, based on the abnormal operation condition of the printing plate determined in step S31, the printing plate defect location processing is then performed. In this step, the defect area appearing in the pattern is analyzed based on the decreasing visual performance of the printed pattern. For example, image processing techniques such as edge detection, texture analysis, morphological operations, etc. are used to further identify specific defects in the pattern. These defects manifest as uneven printing, pattern overlap, color difference, spots, line breaks, etc., and the specific manifestations depend on the condition of the printing plate. By performing multiple filtering processes, sharpening and noise reduction on the pattern image, the printing defect area is highlighted and the precise defect location is obtained. In this process, edge detection algorithms such as Sobel and Canny are used, or image segmentation technology is used to divide the pattern into regions to further determine the spatial position of each defect area. The defect location data obtained contains information such as the coordinates, area, and shape of the defect. This data provides an accurate basis for degradation trend prediction and structural aging assessment in subsequent steps.

[0108] Step S33: predicting the structural degradation trend of the printing plate defect area based on the printing plate defect location data;

[0109] In an embodiment of the present invention, after obtaining printing plate defect location data, this data is used to predict the structural degradation trend of the defective area. Degradation trends reflect the gradual structural damage of certain areas of the printing plate during use due to factors such as repeated printing pressure, temperature fluctuations, or physical wear. In this step, the location data is analyzed to identify the type and extent of the defective area. A degradation trend analysis model is established, which can simulate and predict the degradation behavior of the defective area over time. Specifically, historical data and field test data are combined to analyze the degradation pattern of the defective area using data regression analysis or empirical rules. For example, for cracks on the printing plate surface, the crack growth trend is predicted based on parameters such as crack growth rate, depth, and width. By calculating the degradation rate and defect area change trend, the overall degradation of the printing plate is further inferred. This process combines regular scanning and image recording of the defective area by the data acquisition system to establish a degradation model, thereby obtaining a degradation prediction value for each defective area and generating a degradation trend curve to clearly define the degree of degradation of each defective area and its future development direction.

[0110] Step S34: Evaluate the degree of structural aging of the printing plate according to the structural degradation trend of the printing plate defect area and the printing plate defect location data.

[0111] In this embodiment of the present invention, after determining the degradation trend in step S33, the overall structural aging of the printing plate is further assessed in combination with the defect location data. Structural aging refers to the degradation of the physical properties of a printing plate due to prolonged use, including reduced surface hardness, material fatigue, and decreased printing accuracy. In this step, a printing plate aging assessment model is established based on the degradation trend of the defective area, combined with factors such as the plate's usage time and workload. The model calculates the aging of the printing plate by combining the degradation trend with the location data. For example, for a cracked area, the aging of that area is calculated based on the crack's expansion trend and the area of crack expansion. If the crack area is large and the expansion trend is obvious, it indicates that the area is aging rapidly. This method provides a comprehensive assessment of the aging of the entire printing plate. The specific assessment results are presented in numerical form, resulting in a printing plate structural aging index, which reflects the service life of the entire printing plate and the current aging level. This index can be used to guide subsequent printing plate maintenance, replacement, and improvement efforts.

[0112] Preferably, step S33 includes the following steps:

[0113] Step S331: measuring the printing plate defect size data according to the printing plate defect location data;

[0114] In an embodiment of the present invention, the specific size of each defect is further calculated using the printing plate defect location data obtained in step S32. This process is implemented through image processing technology, in which each defect area is identified using image segmentation technology. For each defect area, the boundary of the defect is accurately delineated by an edge detection algorithm (such as Canny edge detection or Sobel operator). Then, the region growing algorithm is used to further determine the key features such as the area, shape, and boundary contour of the defect area. By calculating the area and geometric features (such as maximum length, minimum width, depth, etc.) of the defect area, the size data of the printing plate defect is obtained. These defect size data provide a basis for subsequent structural wear analysis and degradation trend prediction. During the measurement process, a high-resolution scanner or digital imaging system is used to obtain image data of the printing plate, and ensure that the image is clear and noise-free to avoid defect measurement errors caused by poor image quality. In addition, in order to improve the measurement accuracy, multiple measurements and data averaging are used to further optimize the accuracy of the defect size data.

[0115] Step S332: detecting the wear of the printing plate structure based on the printing plate defect location data;

[0116] In an embodiment of the present invention, structural wear detection of the printing plate is performed based on the defect size data obtained in step S331. In this step, it is necessary to analyze whether the defects are concentrated in specific areas of the printing plate based on the distribution of the defects, and calculate the distribution density of the defective areas. By calculating the morphological changes (e.g., shape irregularity, crack propagation direction, etc.) of each defective area, the wear of the printing plate in these areas can be inferred. Specifically, for defects such as cracks or pits, stress analysis methods are used to quantitatively analyze the wear of the defective areas, combined with the physical properties of the printing plate material (e.g., hardness, wear resistance, etc.). For example, by using finite element analysis (FEA) technology to simulate the stress conditions of the printing plate in the working environment, the stress distribution of each area is calculated, and the degree of wear is inferred based on the location of the defects. In this process, the stress analysis model used will take into account the material properties of the surface and interior of the printing plate to further infer the wear condition of the printing plate. Through this process, the spatial distribution characteristics of the printing plate wear are obtained, providing data support for the stress growth trend prediction in the subsequent steps.

[0117] Step S333: testing the stress growth trend of the printing plate structure based on the wear condition of the printing plate structure;

[0118] In an embodiment of the present invention, based on the wear information obtained in step S332, a stress growth trend test of the printing plate structure is then performed. This step analyzes the accumulation and growth of stress in the printing plate during use by combining the stress-strain theory of materials science. Specifically, finite element analysis (FEA) is used to simulate the workload and external environmental factors (such as temperature changes and mechanical vibration) encountered by the printing plate during long-term use, and the stress distribution of each defective area and the entire printing plate is evaluated. Based on the structural wear, the model dynamically adjusts the stress value of each area to reflect the stress changes caused by wear. Using the stress growth model, stress evolution curves of the printing plate as a whole and in each local area are obtained. These curves reflect how the printing plate material gradually accumulates stress during the wear process under different working conditions. Through this analysis, the stress growth of a specific area of the printing plate at a certain point in the future is predicted, and the relevant stress increase trend data is obtained. These data provide a scientific basis for evaluating the structural aging and crack propagation of the printing plate.

[0119] Step S334: predicting the printing plate crack growth trend based on the printing plate structural stress growth trend and the printing plate defect size data;

[0120] In an embodiment of the present invention, the stress growth trend data obtained in step S333 is combined with the defect size data in step S331 to predict the crack expansion trend. Crack expansion is one of the main manifestations of aging of the printing plate structure. Therefore, accurately predicting the crack expansion trend is crucial for evaluating the service life of the printing plate. In this step, a crack expansion model is used, combined with material mechanics theory, to infer the expansion behavior of the crack under different stress conditions. Specifically, for each defect area, the speed and direction of crack expansion are predicted based on the stress growth trend and defect size data. For example, by calculating the stress concentration level and local stress distribution at the crack tip, the crack expansion trend under different external forces is evaluated. Based on the starting position of the crack, the stress field and the geometric characteristics of the defect, the model predicts the crack expansion path, expansion speed and the impact of the crack on the entire printing plate structure. The prediction results obtained will show the extent of crack expansion at a certain point in the future, guiding subsequent maintenance and replacement decisions.

[0121] Step S335: evaluating the degree of decline in printing accuracy of the printing plate based on the crack growth trend of the printing plate and the stress growth trend of the printing plate structure;

[0122] In an embodiment of the present invention, the degree of decline in printing accuracy of the printing plate is further evaluated based on the crack expansion trend predicted in step S334 and the stress growth trend in step S333. When cracks appear on the printing plate, the cracks will affect the printing accuracy, resulting in phenomena such as offset, missing or unclear patterns during the printing process. By establishing a printing accuracy evaluation model and combining the prediction results of crack expansion, the impact of cracks on printing quality is analyzed. In specific operations, the impact of cracks on pattern alignment and detail restoration during the printing process is evaluated based on their location and expansion trend. By simulating the printing process under different crack expansion degrees and combining the printing accuracy standard, the printing accuracy loss value under each crack state is obtained. This accuracy loss value reflects the degree of decline in printing quality caused by cracks during the printing process. Through this process, the accuracy attenuation of the printing plate can be clearly evaluated, and reference data can be provided for subsequent process adjustments and maintenance.

[0123] Step S336: predicting the structural degradation trend of the defective area of the printing plate based on the degree of reduction in printing accuracy of the printing plate and the trend of crack expansion of the printing plate.

[0124] In an embodiment of the present invention, the structural degradation trend of the defective area of the printing plate is predicted based on the degree of decline in printing accuracy obtained in step S335 and the crack expansion trend in step S334. In this step, the overall degradation trend of the printing plate is further predicted by combining multiple factors of crack expansion and precision decline. The impact of crack expansion in each defective area on the overall printing quality is evaluated, and the expansion rate and degree of degradation of the degraded area are calculated based on the wear and stress distribution. Specifically, for areas where cracks already exist, based on the crack expansion trend and stress growth, it is predicted how the physical properties of the area will gradually decline, leading to degradation of the overall structure of the printing plate. Through multi-factor analysis, the degradation trend of the entire printing plate is derived, providing predictive information on degradation rate, regional expansion, and other aspects. This predictive information provides scientific decision-making support for subsequent maintenance, replacement, and improvement of the printing plate.

[0125] It is particularly important that step S335 includes the following steps:

[0126] Step S3351: detecting changes in physical properties of the printing plate surface based on the crack growth trend of the printing plate;

[0127] In an embodiment of the present invention, according to the crack expansion trend of the printing plate, high-resolution image acquisition is performed on the surface of the printing plate, combined with microscope image analysis technology to detect the crack expansion situation. A high-resolution microscope camera is used to scan the surface of the printing plate to capture surface detail images. By analyzing the expansion trend of the crack in the time series, the position, size and morphological changes of the crack expansion are recorded, and the crack features are extracted using image processing software (such as OpenCV) to identify the specific area of crack expansion. Combined with the position information of the crack expansion area, precision instruments (such as surface roughness measuring instruments) are used to measure the physical properties of the area around the crack, with special attention to parameters such as surface hardness, roughness and elastic modulus. Changes in these physical properties are usually caused by structural changes caused by crack expansion. Therefore, by measuring the changes in physical properties during crack expansion, the degradation of the surface material of the printing plate can be obtained, providing data support for subsequent predictions of ink penetration.

[0128] Step S3352: predicting ink penetration of the printing plate based on changes in physical properties of the printing plate surface and crack growth trends of the printing plate;

[0129] In an embodiment of the present invention, after detecting changes in the physical properties of the printing plate surface, the penetration of the ink is analyzed in combination with the crack expansion trend. The penetration characteristics of the ink on the printing plate surface are derived by correlating the physical property data (such as surface roughness, hardness and elastic modulus) with the crack expansion trend. For the crack expansion area, it usually causes the surface hardness of the area to decrease and the roughness to increase. These changes will directly affect the penetration depth of the ink. The ink permeability is calculated using a physical model that takes into account the surface degradation effect caused by cracks. By establishing a relationship between surface roughness and ink penetration rate, the ink penetration of the crack expansion area is predicted. Combining the printing plate surface property data with the crack expansion trend, a prediction model is constructed to obtain the spatial distribution of the ink penetration degree in each area, and further determine which areas are at risk of uneven ink distribution.

[0130] Step S3353: determining the uneven distribution of ink on the surface of the printing plate according to the ink penetration of the printing plate;

[0131] In an embodiment of the present invention, the spatial distribution data obtained based on the ink penetration situation is used to determine whether the ink distribution on the surface of the printing plate is uniform. By analyzing the depth and speed differences of ink penetration, combined with the influence of the crack propagation area, the uniformity of the surface ink distribution is determined. In actual operation, a high-precision ink sensor or ink volume measuring instrument is used to measure the ink content in different areas of the printing plate surface and compare it with the predicted data. If there are areas with a larger ink penetration depth, the ink distribution will be uneven. This problem affects the quality of the printed product, especially in terms of detail processing and color reproduction. Therefore, by evaluating the ink penetration situation, it is clear whether there is uneven ink distribution on the surface of the printing plate, and further provides a basis for evaluating the decline in printing accuracy.

[0132] Step S3354: estimating the deformation degree of the printing plate based on the stress growth trend of the printing plate structure;

[0133] In an embodiment of the present invention, the degree of deformation of the printing plate is estimated based on the growth trend of the structural stress of the printing plate. Structural stress refers to the cumulative effect of the mechanical load to which the printing plate is subjected during operation. During the printing process, long-term friction, pressure and thermal effects can cause deformation on the surface of the printing plate. The surface of the printing plate is monitored in real time using a stress sensor to collect stress distribution data of the printing plate. By analyzing the changes in stress distribution and combining the stress-deformation theoretical model, the degree of deformation of the printing plate is evaluated. Specifically, in actual operation, the changes in stress on the surface of the printing plate are detected by using a stress analysis instrument (such as a strain gauge). The stress data can reveal the specific trend of the deformation of the printing plate through time series analysis. This data serves as the basis for judging whether the printing plate has undergone significant deformation. A printing plate with a large degree of deformation will directly affect the accuracy and quality of printing.

[0134] Step S3355: detecting unevenness of the printing plate surface according to the degree of deformation of the printing plate;

[0135] In an embodiment of the present invention, based on the deformation data obtained in the previous step, a high-precision surface detection device such as a three-dimensional scanner is used to detect the unevenness of the surface of the printing plate. By processing the three-dimensional point cloud data of the surface of the printing plate, it is possible to clearly identify whether the printing plate has surface unevenness, especially the deformation in the stress-bearing area. By analyzing the degree of surface unevenness and combining it with the deformation data obtained previously, the actual state of the printing plate can be further confirmed. For example, deformation in certain areas causes the surface to be concave or convex, which affects the uniform distribution of ink and the printing effect. Therefore, the detection of surface unevenness provides an important basis for the subsequent evaluation of printing accuracy.

[0136] Step S336: Evaluate the degree of reduction in printing accuracy of the printing plate based on the unevenness of the printing plate surface and the uneven distribution of ink on the printing plate surface.

[0137] In an embodiment of the present invention, the degree of reduction in printing accuracy of a printing plate is evaluated by comprehensively considering the surface unevenness and the uneven distribution of ink. The data on surface unevenness and the data on uneven distribution of ink are correlated and analyzed to identify which uneven areas also have uneven ink penetration. These areas typically lead to a reduction in the quality of printed products, especially in printing processes requiring high precision. Then, based on the surface unevenness and uneven distribution of ink, a weighted algorithm is used to calculate the degree of reduction in printing accuracy of each area, and the percentage of reduction in printing accuracy of the entire printing plate is obtained, thereby providing guidance for subsequent optimization or replacement of the printing plate.

[0138] It is particularly important that step S34 includes the following steps:

[0139] Step S341: detecting the degree of friction growth of the printing plate according to the structural degradation trend of the printing plate defect area and the printing plate defect location data;

[0140] In an embodiment of the present invention, image data of defective areas on the surface of a printing plate is acquired using high-precision image acquisition technology (such as a microscope or scanning electron microscope). Image processing software (such as OpenCV) is used to analyze the images and extract the specific location, size, and morphology of the defective area. Combined with this data, the structural degradation trend of the area is further analyzed. Specifically, by comparing images of the printing plate acquired at different times, the expansion trend of the defective area can be identified, and the degree of structural degradation in that area can be deduced. After acquiring the location data of the defective area, real-time monitoring is performed in conjunction with a friction sensor (such as a surface stress sensor or a friction coefficient sensor). The friction sensor can measure the changes in friction generated by the printing plate surface during operation, particularly in high-friction areas. Through long-term data acquisition and recording, combined with the degradation trend of the defective area, the friction growth of specific areas of the printing plate can be analyzed. The rate and degree of friction growth are closely related to the degree of structural degradation, so the friction data is used to infer the degree of aging of the defective area. During this process, by continuously tracking the relationship between the friction data and the structural degradation trend, the specific degree of friction growth is determined, and the friction growth rate of the printing plate is derived. This process predicts the future trend of friction growth by setting a threshold or using a data fitting model.

[0141] Step S342: determining the thermal expansion degree of the printing plate based on the friction growth degree of the printing plate;

[0142] In an embodiment of the present invention, the degree of thermal expansion of the printing plate is further deduced from friction growth data. Increased friction is often accompanied by a local temperature increase. In particular, during the printing process, heat generated by the printing plate in the friction zone can lead to thermal expansion. To accurately measure the degree of thermal expansion, a high-precision temperature sensor (such as an infrared thermal imager or thermocouple) is required to monitor the temperature changes on the printing plate surface in real time. Based on the relationship between friction growth data and surface temperature data, the degree of thermal expansion of the printing plate is calculated using a theoretical thermal expansion model. Specifically, the thermal expansion change of the material is estimated based on the rate of temperature increase of the printing plate surface and the friction data, using the coefficient of thermal expansion (determined by the properties of the printing plate material). Generally, when friction increases, the temperature of a local area rises, causing the volume of that area to expand. By calculating the degree of thermal expansion in that area, the impact of thermal effects on the deformation of the printing plate surface can be determined. In this step, combined analysis of the temperature distribution map and the friction distribution map further verifies the relationship between thermal expansion and friction growth, and accurately calculates the deformation data caused by thermal expansion.

[0143] Step S343: predicting the growth trend of the defective region area gradient according to the printing plate friction growth degree and the printing plate friction growth degree;

[0144] In an embodiment of the present invention, after obtaining the degree of friction growth of the printing plate, the growth trend of the area of the defective area is predicted in combination with the friction data of the defective area. By comparing the friction data and structural degradation trends of different areas, a correlation model between friction and defect area growth is established. By monitoring the friction data for a long time, the trend of friction change over time is obtained. Then, the friction growth data is combined with the expansion trend of the defective area, and the gradient growth of the defective area is predicted by the data fitting method. This prediction process is based on the physical relationship between friction and defect expansion. An increase in friction usually means that the surface damage of the local area is aggravated, which in turn leads to the expansion of the defective area. Therefore, by collecting friction data in real time, the future expansion range and expansion speed of these defective areas are predicted. Specifically, data processing techniques such as regression analysis and trend analysis are used, combined with known friction growth data, to deduce the growth trend of the defective area.

[0145] Step S344: Estimate the aging degree of the printing plate structure according to the gradient growth trend of the defective region area.

[0146] In an embodiment of the present invention, the degree of structural aging of the printing plate is further estimated based on the gradient growth trend of the defect area area. Structural aging is mainly manifested in material degradation, surface damage and functional decline. By combining the gradient growth trend of the defect area area, a relationship model between defect growth and structural aging is established. Specifically, the area growth rate of the defect area reflects the degree of aging of the surface material of the printing plate. To achieve this, a surface roughness instrument (such as an atomic force microscope) is used to carefully scan the surface of the printing plate to obtain structural change data of the defect area. Then, by analyzing the correlation between the area change of the defect area and physical properties such as surface hardness and elastic modulus, the degree of aging of the printing plate is further calculated. By analyzing the area growth data of multiple defect areas, the gradient growth trend between regions is constructed, and the overall aging of the printing plate is calculated. This process involves a comprehensive evaluation of multiple parameters, such as the rate of crack propagation, the increase in surface friction, temperature changes, etc. Combining these data, the degree of structural aging of the printing plate can be accurately estimated.

[0147] Preferably, step S4 includes the following steps:

[0148] Step S41: predicting the growth trend of the progressive error of the printing plate according to the aging degree of the printing plate structure;

[0149] In an embodiment of the present invention, the structural aging of a printing plate is analyzed to predict its progressive error growth trend. Based on the structural wear and crack growth data obtained in previous steps (e.g., steps S33 and S332), and combined with the printing plate's usage history and operating environment (e.g., temperature, humidity, mechanical load, etc.), a structural aging model is constructed. Using this data, a stress analysis method (e.g., finite element analysis) is used to simulate the stress distribution and progressive wear of the printing plate during continued use. The structural aging model comprehensively analyzes the printing plate, taking into account the wear rate, crack growth path, and physical properties of the printing plate (e.g., hardness, elastic modulus, etc.) in different regions. Specifically, experimental data is collected, including multiple load tests under different conditions, and the parameters of the aging model are adjusted based on the experimental results. Through long-term aging experiments, the progressive error growth trend of different regions of the printing plate is determined. For example, as the printing plate ages, cracks gradually expand, affecting printing accuracy and leading to a cumulative increase in progressive error. The predicted error growth trend data can be used as a basis for determining printing anomalies in subsequent steps.

[0150] Step S42: determining the printing abnormality of the printing plate according to the growth trend of the progressive error of the printing plate;

[0151] In an embodiment of the present invention, after obtaining the progressive error growth trend of the printing plate, a judgment of printing abnormalities is performed. This step mainly compares and analyzes the progressive error with the working status of the printing plate to detect whether a printing abnormality occurs. By measuring and recording the actual error value of each batch of printing and comparing it with the standard error in the initial state, a gradually increasing error value is obtained. If the error value exceeds the set allowable range, it can be considered that the printing plate has an abnormality. In order to accurately judge the abnormality, multiple error thresholds are set, and classification management is performed for different error sizes. For example, if the error exceeds a certain critical value, the printed pattern will be blurred or displaced, indicating a serious printing abnormality. By continuously tracking and recording changes in the error trend, it is determined when the abnormality occurred and the severity of the abnormality is evaluated. This process needs to be combined with continuous monitoring of printing quality in actual production, and the finished product patterns of each round of printing are carefully compared with image processing tools (such as images taken by high-precision cameras) to quantify errors and identify abnormalities in a timely manner.

[0152] Step S43: predicting the failure state of the printing plate structure performance based on the printing abnormality of the printing plate and the growth trend of the progressive error of the printing plate;

[0153] In an embodiment of the present invention, this step further predicts the structural performance failure state of the printing plate by combining the progressive error growth trend of the printing plate and printing anomalies. In the previous step, data on the error changes and the gradual expansion of cracks generated during the use of the printing plate were obtained. Using this data, combined with material mechanics and fatigue analysis theory, a failure prediction model was established. Specifically, the stress distribution and defect expansion of the printing plate at different stages were calculated to analyze their impact on printing performance. By superimposing the known error growth trend with the crack expansion trend, it was predicted that at a certain point in time, the printing plate would experience irreversible performance loss due to structural aging and further crack expansion. Specifically, by simulating the effects of various external stresses (such as shock and vibration) on the printing plate, combined with the gradually increasing error, the timing and failure mode of the failure state were predicted. For example, further crack expansion could lead to complete failure in a localized area, resulting in inaccurate printing of the pattern, thereby affecting the printing quality of the entire production line. The purpose of this step is to predict the point of structural performance failure by analyzing the gradual accumulation of aging and damage, and to provide data support for subsequent preventive measures.

[0154] Step S44: determining the printing abnormality defect according to the failure state of the printing plate structure performance and the decreasing trend of the visual performance of the printed pattern;

[0155] In an embodiment of the present invention, after obtaining the structural performance failure state of the printing plate in step S43, the specific situation of the abnormal printing defect is determined in combination with the decreasing trend of the visual performance of the printed pattern, and a high-definition image of the printed product is obtained through image acquisition technology, and a detailed analysis is performed to record the changes in the visual effect of the pattern during each printing process. By comparing the deviation between the actual pattern and the standard pattern, the error type in the printing process, such as position offset, blur, color difference, etc., is quantified. In particular, as the printing plate ages and the structural performance fails, the visual characteristics of the pattern such as clarity and alignment will gradually deteriorate. Therefore, for each round of printing process, an image is obtained by a high-precision camera system or scanning system, and compared with the standard pattern, and the degree of deviation of the pattern is detected using image recognition technology (such as edge detection, feature matching, etc.). In this step, it is also necessary to combine historical data and aging models to analyze the trend of the decreasing visual performance of the pattern, evaluate the changes in the visual effect during each printing process, and thus accurately identify abnormal printing defects. This analysis result will help identify and classify abnormal defects and ensure the accurate positioning of printing quality problems.

[0156] Step S45: Optimizing the printing defect detection model through model training using the printing abnormality defect situation to obtain an optimized training model for printing defect detection.

[0157] In an embodiment of the present invention, after identifying abnormal printing defects in step S44, the printing defect detection model is optimized and trained using these defect data. In this step, the abnormal defect conditions and pattern deviation data detected in the previous step are used as a training set and input into the defect detection model. This model is a defect detection algorithm based on image recognition technology (such as deep learning, convolutional neural networks, etc.), which is specifically used to identify and classify printing defects. By inputting a large amount of abnormal data, the model performs backpropagation and parameter adjustment to optimize its detection accuracy. During the training process, the data used includes different types of defect patterns (such as cracks, blur, displacement, etc.), and labels are classified according to the severity of the defects. Through continuous optimization, a detection model that can accurately detect and classify various printing defects is obtained. The optimized model can perform real-time detection of printing defects in future production and provide timely feedback to ensure printing quality. Data enhancement technology is also introduced in the model training process to further improve the robustness and accuracy of the model by rotating, cropping, transforming and other operations on the defect pattern.

[0158] The present invention also provides a training system for a printing defect detection model, which is used to execute the training method for a printing defect detection model as described above. The training system for a printing defect detection model includes:

[0159] A model building module is used to obtain appearance data of a printed product; perform image acquisition of the printed product on the appearance data of the printed product to obtain image data of the printed product; perform image optimization processing on the image data of the printed product to obtain optimized image data of the printed product; and construct a printing defect detection model based on the optimized image data of the printed product to obtain a printing defect detection model;

[0160] A visual performance degradation trend assessment module is used to determine the abnormal distortion state of the printed pattern based on the optimized data of the printed product image; detect the gradient blurring state of the printed pattern details based on the optimized data of the printed product image; and assess the visual performance degradation trend of the printed pattern based on the gradient blurring state of the printed pattern details and the abnormal distortion state of the printed pattern;

[0161] A structural aging assessment module is used to determine abnormal operation conditions of the printing plate based on the decreasing trend of visual performance of the printed pattern; to perform printing plate defect location processing based on the abnormal operation conditions of the printing plate to obtain printing plate defect location data; and to assess the structural aging degree of the printing plate based on the printing plate defect location data;

[0162] The model training optimization module is used to predict the growth trend of the progressive error of the printing plate according to the aging degree of the printing plate structure; determine the abnormal printing defects according to the growth trend of the progressive error of the printing plate and the decreasing trend of the visual performance of the printed pattern; use the abnormal printing defects to optimize the model training of the printing defect detection model to obtain the printing defect detection optimization training model.

[0163] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A training method for a printing defect detection model, characterized in that: The following steps are involved: Step S1: Acquire appearance data of a printed product; perform image capture of the printed product on the appearance data of the printed product to obtain image data of the printed product; Performing image optimization processing on the image data of the printed product to obtain optimized image data of the printed product; A printing defect detection model is constructed based on the optimized data of the printed product image to obtain a printing defect detection model; Step S2: determining the abnormal state of printed pattern distortion based on the optimized data of the printed product image; Detect the gradient blur of printed pattern details based on the optimized data of the printed finished product image; Evaluate the decreasing trend of printed pattern visual performance based on the gradient blur of printed pattern details and the abnormal distortion of printed pattern; Step S3: determining an abnormal operation condition of the printing plate based on the decreasing trend of the visual performance of the printed pattern; Utilize the abnormal operation condition of the printing plate to carry out printing plate defect location processing and obtain printing plate defect location data; Evaluate the degree of structural aging of printing plates based on printing plate defect location data; Step S4: predicting the growth trend of the progressive error of the printing plate according to the aging degree of the printing plate structure; Determine printing abnormalities and defects based on the growth trend of progressive errors in the printing plate and the decreasing trend of visual performance of the printed pattern; The printing defect detection model is trained and optimized using the printing abnormal defect situation to obtain the printing defect detection optimization training model.

2. The training method for a printing defect detection model according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: setting the resolution of the visible light camera to 1920×1080 pixels, the minimum light sensitivity to 0.01 lux, and the frame rate to 30 FPS; Step S12: Obtaining appearance data of the printed product; Step S13: collecting the printed product image data based on the appearance data of the printed product using the visible light camera to obtain the printed product image data; Step S14: performing image optimization processing on the printed product image data to obtain the printed product image optimized data; Step S15: constructing a printing defect detection model based on the optimized data of the printed product image to obtain a printing defect detection model.

3. The training method for a printing defect detection model according to claim 2, characterized in that: Step S14 includes the following steps: Step S141: performing a gradient amplitude distribution analysis on the printed product image data to obtain a gradient amplitude distribution of the printed product image; Step S142: evaluating the clarity of the printed product image based on the gradient amplitude distribution of the printed product image being greater than 60; Step S143: performing local texture intensity recognition of the printed product image based on the printed product image data to obtain local texture intensity data of the image; Step S144: evaluating the initial quality of the printed product image based on the image local texture intensity data and the clarity of the printed product image; Step S145: Dividing the initial quality of the printed product image into 2200 grid regions to obtain grid region data of the printed product image quality; Step S146: detecting the high-frequency component loss area of the finished product image based on the quality grid area data of the printed product image; Step S147: performing image optimization processing on the high-frequency component missing area of the finished product image to obtain optimized data of the printed finished product image.

4. The training method for a printing defect detection model according to claim 1, characterized in that: The determination of the abnormal state of the printed pattern distortion in step S2 includes: Optimize data collection based on the printed product image to obtain the external contour of the printed product pattern; Measure the overall proportion change of the printed product pattern based on the printed product image optimization data; Integrate the external contour of the printed product pattern and the overall proportion change of the printed product pattern to obtain the characteristic data of the printed product pattern; Acquire printed image structure design data, and perform image registration processing on printed product pattern feature data based on the printed image structure design data to obtain printed product image registration data; Measure the node offset of the printed pattern based on the image registration data of the printed product; Evaluate the abnormal aspect ratio deviation of the image based on the node deviation of the printed pattern; Detect the stretch-compression degree of the printed pattern based on the abnormal deviation of the image aspect ratio; The abnormal state of printed pattern distortion is determined based on the degree of stretching and compression of the printed pattern and the abnormal deviation of the image aspect ratio.

5. The training method for a printing defect detection model according to claim 1, characterized in that: The detection of gradient blur of printed pattern details in step S2 includes: Collect the texture features of the printed product image based on the optimized data of the printed product image; Identify the energy attenuation trend of local texture in the printed product image based on its texture features; The degree of detail loss of printed patterns is evaluated based on the energy attenuation trend of local texture in the image; Detect the continuity interruption of the printed pattern based on the degree of loss of printed pattern details and the trend of local texture energy attenuation in the image; Identify the image texture direction vector offset based on the texture features of the printed product image; Determine the weakening of pattern details in the printed product based on the image texture direction vector offset and the degree of loss of printed pattern details; The gradient blurring of printed pattern details is detected based on the weakening of printed pattern details and the interruption of printed pattern continuity.

6. The training method for a printing defect detection model according to claim 1, characterized in that: The evaluation of the visual performance degradation of the printed pattern in step S2 includes: Detect the structural distortion of the printed pattern based on the gradient blur of the printed pattern details and the abnormal distortion of the printed pattern; Determine the increasing difficulty trend of printed pattern structure recognition based on the printed pattern structure distortion; Predict the sharp decline in the decoding rate of printed patterns based on the increasing trend of printed pattern structure recognition difficulty; Detecting printed pattern information reading failures based on the sharp decline in printed pattern decoding rates and the increasing difficulty in identifying printed pattern structures; Predict the breakage of the printed pattern traceability chain based on the failure of printed pattern information reading and the sharp decline of printed pattern decoding rate; The decreasing trend of the visual performance of printed patterns is evaluated based on the breaks in the printed pattern traceability chain and the failure in reading the printed pattern information.

7. The training method for a printing defect detection model according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: determining an abnormal operation condition of the printing plate based on the decreasing trend of the visual performance of the printed pattern; Step S32: performing printing plate defect location processing using the printing plate operation abnormality to obtain printing plate defect location data; Step S33: predicting the structural degradation trend of the printing plate defect area based on the printing plate defect location data; Step S34: Evaluate the degree of structural aging of the printing plate according to the structural degradation trend of the printing plate defect area and the printing plate defect location data.

8. The training method for a printing defect detection model according to claim 7, characterized in that: Step S33 includes the following steps: Step S331: measuring the printing plate defect size data according to the printing plate defect location data; Step S332: detecting the wear of the printing plate structure based on the printing plate defect location data; Step S333: testing the stress growth trend of the printing plate structure based on the wear condition of the printing plate structure; Step S334: predicting the printing plate crack growth trend based on the printing plate structural stress growth trend and the printing plate defect size data; Step S335: evaluating the degree of decline in printing accuracy of the printing plate based on the crack growth trend of the printing plate and the stress growth trend of the printing plate structure; Step S336: predicting the structural degradation trend of the defective area of the printing plate based on the degree of reduction in printing accuracy of the printing plate and the trend of crack expansion of the printing plate.

9. The training method for a printing defect detection model according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: predicting the growth trend of the progressive error of the printing plate according to the aging degree of the printing plate structure; Step S42: determining the printing abnormality of the printing plate according to the growth trend of the progressive error of the printing plate; Step S43: predicting the failure state of the printing plate structure performance based on the printing abnormality of the printing plate and the growth trend of the progressive error of the printing plate; Step S44: determining the printing abnormality defect according to the failure state of the printing plate structure performance and the decreasing trend of the visual performance of the printed pattern; Step S45: Optimizing the printing defect detection model through model training using the printing abnormality defect situation to obtain an optimized training model for printing defect detection.

10. A training system for a printing defect detection model, characterized in that: A system for executing the training method for a printing defect detection model according to claim 1, wherein the training system for a printing defect detection model comprises: A model building module is used to obtain appearance data of a printed product; perform image acquisition of the printed product on the appearance data of the printed product to obtain image data of the printed product; perform image optimization processing on the image data of the printed product to obtain optimized image data of the printed product; and construct a printing defect detection model based on the optimized image data of the printed product to obtain a printing defect detection model; A visual performance degradation trend assessment module is used to determine the abnormal distortion state of the printed pattern based on the optimized data of the printed product image; detect the gradient blurring state of the printed pattern details based on the optimized data of the printed product image; and assess the visual performance degradation trend of the printed pattern based on the gradient blurring state of the printed pattern details and the abnormal distortion state of the printed pattern; A structural aging assessment module is used to determine abnormal operation conditions of the printing plate based on the decreasing trend of visual performance of the printed pattern; to perform printing plate defect location processing based on the abnormal operation conditions of the printing plate to obtain printing plate defect location data; and to assess the structural aging degree of the printing plate based on the printing plate defect location data; The model training optimization module is used to predict the growth trend of the progressive error of the printing plate according to the aging degree of the printing plate structure; determine the abnormal printing defects according to the growth trend of the progressive error of the printing plate and the decreasing trend of the visual performance of the printed pattern; use the abnormal printing defects to optimize the model training of the printing defect detection model to obtain the printing defect detection optimization training model.

Images