High-precision printing quality detection method and system

Abstract

The invention discloses a high-precision printing quality detection method and system, and relates to the technical field of image processing. The method comprises the following steps: configuring an image sensor for a printing quality inspection area, and performing sensing acquisition on a target printing product to obtain a printing image; the printing image is returned, image segmentation and gray processing are carried out according to the boundary of the printing product, and a gray printing image is determined; a standard printing template is obtained, a pixel stack space is constructed, and a printing registration module is supervised and trained; the gray printing image is transmitted to a printing registration module, and defect pixel features are determined through pixel transition sampling registration and pixel interlaced sampling registration; and traversing defect pixel features, performing defect type matching based on a defect matcher, and outputting printing quality data. The technical problem that in the prior art, printing quality detection precision and efficiency are low is solved, and the technical effect of improving the printing quality detection precision and efficiency is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a high-precision printing quality detection method and system. Background Art

[0002] With the continuous development of modern industrial manufacturing, especially the continuous upgrading and expansion of the printing industry, printing quality detection has become a vital part of the production process. Traditional printing quality detection methods mainly rely on manual visual inspection or simple mechanical detection equipment. These methods have certain limitations in accuracy and efficiency, especially when facing large-scale, high-precision printing needs, they often cannot meet the requirements of fast, accurate, and error-free. In addition, the existing automated detection systems also face the problems of low detection accuracy and insufficient defect recognition capabilities, especially when dealing with complex printed images or minor defects, which are prone to missed detections and false detections. Therefore, how to improve the accuracy and efficiency of printing quality detection, especially in high-speed, large-scale production, to achieve automated and intelligent quality detection, has become a key issue that needs to be urgently solved in the current printing industry. Summary of the invention

[0003] The present application provides a high-precision printing quality detection method and system, which solves the technical problems of low printing quality detection accuracy and efficiency in the prior art.

[0004] In view of the above problems, the present application provides a high-precision printing quality detection method and system.

[0005] In a first aspect of the present application, a high-precision printing quality detection method is provided, the method comprising: An image sensor is configured in the printing quality inspection area, and the target printing product is sensed and collected to obtain a printing image; the printing image is transmitted back, and image segmentation and grayscale processing are performed based on the printed product boundary to determine the grayscale printing image; a standard printing template is obtained, and a pixel stack space is constructed to supervise the training of a printing registration module, wherein the printing registration module includes a first pixel registration unit and a second pixel registration unit, and a first acquisition registration is performed based on pixel transition sampling, and a second acquisition registration is performed based on pixel interval sampling; the grayscale printing image is transmitted to the printing registration module, and defective pixel features are determined by performing pixel transition sampling registration and pixel interval sampling registration; the defective pixel features are traversed, defect type matching is performed based on a defect matcher, and printing quality data is output.

[0006] A second aspect of the present application provides a high-precision printing quality detection system, the system comprising: Image acquisition unit: configure image sensors in the printing quality inspection area, perform sensing and acquisition on the target printing product to obtain the printing image; image processing unit: transmit the printing image back, perform image segmentation and grayscale processing based on the printing product boundary, and determine the grayscale printing image; training unit: obtain the standard printing template, and construct a pixel stack space, supervise the training of the printing registration module, wherein the printing registration module includes a first pixel registration unit and a second pixel registration unit, and performs the first acquisition registration by pixel transition sampling, and performs the second acquisition registration by pixel interval sampling; registration unit: transmit the grayscale printing image to the printing registration module, and determine the defective pixel features by performing pixel transition sampling registration and pixel interval sampling registration; defect matching unit: traverse the defective pixel features, perform defect type matching based on the defect matcher, and output printing quality data.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, an image sensor is configured in the printing quality inspection area, and the target printing product is sensed and collected to obtain a printing image. Then, the printing image is transmitted back, and the image is segmented and grayscale processed based on the boundary of the printing product to determine the grayscale printing image. Further, a standard printing template is obtained, and a pixel stack space is constructed to supervise the training of the printing registration module, wherein the printing registration module includes a first pixel registration unit and a second pixel registration unit, and the first acquisition registration is performed by pixel transition sampling, and the second acquisition registration is performed by pixel interval sampling. Then, the grayscale printing image is transmitted to the printing registration module, and the defective pixel features are determined by performing pixel transition sampling registration and pixel interval sampling registration. Finally, the defective pixel features are traversed, and defect type matching is performed based on the defect matcher, and the printing quality data is output. The technical problem of low printing quality detection accuracy and efficiency in the prior art is solved, and the technical effect of improving the printing quality detection accuracy and efficiency is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0009] Figure 1 A schematic diagram of a high-precision printing quality detection method provided in an embodiment of the present application; Figure 2 A schematic diagram of the structure of a high-precision printing quality inspection system provided in an embodiment of the present application.

[0010] Explanation of the reference numerals: image acquisition unit 11 , image processing unit 12 , training unit 13 , registration unit 14 , defect matching unit 15 . DETAILED DESCRIPTION

[0011] The present application solves the technical problems of low printing quality detection accuracy and efficiency in the prior art by providing a high-precision printing quality detection method and system.

[0012] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0013] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0014] Embodiment 1, as Figure 1 As shown, the present application provides a high-precision printing quality detection method, wherein the method comprises: An image sensor is configured in the printing quality inspection area to perform sensing and acquisition on the target printed product to obtain the printed image.

[0015] In the printing quality inspection area (the quality inspection position of the printing production line), according to the size, material and transmission speed of the printed product, a high-resolution and high-sensitivity image sensor, such as an industrial camera or a high-resolution camera, is reasonably configured to ensure that the entire inspection area of ​​the printed product is covered. When the target printed product passes through the printing quality inspection area, the image sensor will be triggered to collect images and obtain the printed image.

[0016] The printed image is transmitted back, and image segmentation and grayscale processing are performed based on the printed product boundary to determine a grayscale printed image.

[0017] The printed images collected by the image sensor are transmitted to the processing system, and these images are segmented and grayscale processed. Specifically, edge detection algorithms (such as Canny algorithm or Sobel algorithm) are used to extract the edge contour information of the printed product, and morphological processing (such as dilation and corrosion operations) is combined to optimize the edge results and remove noise in the image; then the contour extraction algorithm is used to identify the closed boundary area, and the image area related to the printed product is cut out according to the size and position of the boundary area, and the irrelevant background part is removed. After completing the boundary segmentation, the extracted printed image area is grayscaled, and the pixel weighted average method is used to convert each pixel value of the RGB color image into a grayscale value. The calculation formula is: grayscale value = 0.299×R+0.587×G+0.114×B, where R, G, and B are the pixel values ​​of the red, green, and blue channels respectively, and finally a grayscale printed image is obtained.

[0018] A standard printing template is obtained, and a pixel stack space is constructed to supervise the training of a printing registration module, wherein the printing registration module includes a first pixel registration unit and a second pixel registration unit, and a first acquisition registration is performed by pixel transition sampling, and a second acquisition registration is performed by pixel interval sampling.

[0019] The preset standard printing template is loaded from the system, and the standard printing template is used as the ideal benchmark for the target printing product. The pixel information of the standard printing template is used to build a pixel stack space. The pixel stack space is a multidimensional array structure used to store the pixel points and mapping relationship between the standard printing template and the target printing image, and supports subsequent registration and defect detection operations. In the process of building the pixel stack space, the printing registration module is supervised and trained by the annotated standard template data to ensure that it can accurately perform image registration. The printing registration module contains two main units: the first pixel registration unit and the second pixel registration unit; the first pixel registration unit adopts the pixel transition sampling method, and by setting a fixed sampling step (such as sampling every 5 pixels), it quickly extracts the large-scale features of the template image and the target image, and completes the initial rough registration; the second pixel registration unit finely optimizes the preliminary registration result generated by the first pixel registration unit, and uses the pixel interval sampling method (such as sampling every 2 pixels) to align the key pixel area with higher precision in a smaller range.

[0020] The grayscale printing image is transmitted to the printing registration module, and the defective pixel features are determined by performing pixel transition sampling registration and pixel interval sampling registration.

[0021] The grayscale printing image that has been grayed is transmitted to the printing registration module for registration processing. The printing registration module performs pixel transition sampling registration and pixel interval sampling registration through the first pixel registration unit and the second pixel registration unit. Specifically, in the printing registration module, the first pixel registration unit is started, and the pixel transition sampling method is adopted, that is, the pixels of the grayscale printing image are analyzed in a jump manner according to the set sampling step (for example, sampling every 5 pixels), and a preliminary match is made with the large-scale pixel features of the standard printing template; the second pixel registration unit is started for fine processing of the preliminary registration result generated by the transition sampling registration, and the key areas and edge details in the grayscale image are further aligned point by point through the pixel interval sampling method (for example, sampling every 2 pixels); after completing the two-stage registration, the registration results are analyzed for differences, and the abnormal pixels are extracted by calculating the difference in grayscale values ​​or the deviation of template matching, and marked as defective pixel features.

[0022] Furthermore, the pixel transition sampling registration includes: The grayscale printed image is transmitted to the first pixel registration unit, and image pixel scanning and pixel transition recognition are performed according to a preset pixel scanning method to locate pixel transition coordinates; according to a preset window scale, window pixel vectors are collected for the pixel transition coordinates, added into the pixel stack space, and a first pixel matrix is ​​determined; according to the standard printing template, a first standard pixel matrix is ​​determined, and the first standard pixel matrix is ​​registered and proofread with the first pixel matrix to locate abnormal pixel areas, wherein the abnormal pixel areas are unregistered window framed areas.

[0023] Specifically, the grayscale printed image is transmitted to the first pixel registration unit. After the first pixel registration unit is started, the entire grayscale printed image is scanned line by line according to a preset pixel scanning method (such as linear scanning or serpentine scanning). During the scanning process, the grayscale value difference between adjacent pixels is calculated to identify the pixel points whose grayscale value changes exceed the preset threshold, and their position coordinates are recorded. These coordinates are the pixel transition coordinates. Then, with each pixel transition coordinate as the center, according to the preset window scale (such as 3×3, 5×5, etc.), the pixel grayscale vectors around the coordinate are collected to form a window pixel vector, which is used to represent the feature distribution of the local area. These window pixel vectors are stored one by one in the pre-constructed pixel stack space, and finally the first pixel matrix of the target image is generated. Subsequently, the corresponding window pixel vectors are extracted from the standard printing template according to the same pixel transition sampling method and window scale, and the first standard pixel matrix is ​​constructed to ensure that the structures of the two matrices are consistent; the window vectors of the first standard pixel matrix and the first pixel matrix are aligned and proofread one by one, and the error calculation method (such as Euclidean distance) is used to evaluate the matching situation. The formula is in, and are the kth pixel values ​​of the window vector corresponding to the target image and the standard template, respectively, and D is the matching error; for windows whose registration error D exceeds the preset threshold, they are marked as abnormal pixel areas, and the specific coordinates and window range of the abnormal areas are recorded and saved in the abnormal data list.

[0024] Furthermore, the registration and proofreading uses the pixel position as the first registration target and the grayscale vector as the second registration target, wherein the grayscale vector includes the grayscale value of the pixel transition coordinates and the window grayscale gradient feature, and the window grayscale gradient feature includes the gradient direction and the gradient value; if the abnormal pixel area is empty, a qualified quality result is output; if the abnormal pixel area is not empty, the abnormal pixel area is transferred to the second pixel registration unit.

[0025] Preferably, after the grayscale printed image is input into the first pixel registration unit, the registration and proofreading process first takes the pixel position as the first registration target, and aligns the pixel transition coordinates in the target image and the standard template one by one; the system detects the position offset of the pixel transition point according to the preset tolerance range, and if the difference between the actual coordinates of a certain transition point and the corresponding coordinates of the standard template exceeds the tolerance range, the point is marked as a potential abnormal pixel. Subsequently, the grayscale vector is taken as the second registration target, and the grayscale value of each pixel transition coordinate and the window grayscale gradient feature are compared. The grayscale vector consists of two parts: the grayscale value of the transition coordinate and the window grayscale gradient feature. First, the grayscale value difference between the transition point in the target image and the standard template is calculated. If the grayscale value difference exceeds the preset threshold, it is marked as abnormal; then the direction and value of the grayscale gradient in the window are calculated, the gradient direction and gradient value are extracted, and the gradient direction and value are compared with the gradient direction and value of the window corresponding to the standard template. If the deviation exceeds the set threshold, the point is further marked as abnormal. After completing the dual registration and calibration of position and grayscale vectors, the system generates the detection results of the abnormal pixel area: if the abnormal pixel area is empty, it means that all transition coordinates are successfully registered and no abnormalities are detected, then the system directly outputs qualified quality results and records the detection status as qualified; if the abnormal pixel area is not empty, the marked abnormal area data, including the coordinates of the abnormal points, grayscale value differences, gradient features and window ranges, are packaged into flow data and passed to the second pixel registration unit for further analysis, while recording the processing status of the abnormal area.

[0026] Furthermore, pixel-by-pixel sampling and registration are performed, including: The abnormal pixel area is transferred to the second pixel registration unit to collect alternate pixel points and add them into the pixel stack space to generate a second pixel matrix; according to the standard printing template, a second standard pixel matrix is ​​determined; the second pixel matrix and the second standard pixel matrix are calibrated to determine pixel defect features, wherein the defective pixel features include geometric features and grayscale distribution features.

[0027] After the first pixel registration unit completes the preliminary registration and generates the abnormal pixel area, the marked abnormal pixel area is transferred to the second pixel registration unit. After the second pixel registration unit is started, more precise pixel acquisition is performed on the abnormal area based on the alternate point sampling method. Specifically, in each abnormal pixel area, according to the preset alternate sampling strategy (such as sampling every one or two pixels), the higher-resolution pixel information in the abnormal area is extracted, and these collected data are added to the pixel stack space to construct a second pixel matrix representing the characteristics of the abnormal area; the template data corresponding to the abnormal area is extracted from the standard printing template, and the second standard pixel matrix is ​​generated in the same alternate sampling method to ensure that the two matrices are consistent in sampling method and structure; the second pixel matrix and the second standard pixel matrix are proofread point by point, firstly, the differences in geometric features between the two are compared, including the integrity of the contour, the matching degree of the edge shape and the angle offset, and the presence of geometric anomalies is determined by calculating the contour matching rate and the geometric deviation value; secondly, the grayscale distribution characteristics are finely analyzed, and the color difference, texture unevenness or grayscale abnormal points in the abnormal area are identified by comparing the grayscale value change trend, gradient direction and grayscale distribution uniformity of each pixel point; the matching of grayscale distribution features is quantified by calculating the grayscale difference (such as Euclidean distance or correlation coefficient) and the deviation angle of the gradient direction. If a point exceeding the preset threshold is found, it is recorded as a defective pixel feature. After the proofreading is completed, the system generates a detailed description of the defective pixel features, including specific data on its geometric feature abnormalities (such as shape distortion and edge breakage) and grayscale distribution feature abnormalities (such as color deviation and uneven texture).

[0028] The defective pixel features are traversed, defect type matching is performed based on a defect matcher, and printing quality data is output.

[0029] The system traverses and analyzes the marked defective pixel features, and uses a defect matcher (e.g., based on feature templates or machine learning algorithms) to classify these defects into specific defect types (such as color deviation, stains, missing, etc.). Ultimately, the system outputs a set of printing quality data, including detailed information on the inspection results and defect types, for production quality assessment and improvement.

[0030] Furthermore, defect type matching is performed based on the defect matcher, and printing quality data is output, including: A defect matcher is constructed, wherein the defect matcher includes a first matching layer based on a defect state and a second matching layer based on a printing defect type; based on the first matching layer, the defective pixel features are identified and the geometric features are matched to determine a first defect state; based on the second matching layer, the grayscale distribution features are matched according to inter-layer mapping to determine the defect type; and the defective pixel features are combined with the defect type to determine the printing quality data.

[0031] After the defect pixel feature extraction is completed, the system starts the defect matcher to match the defect type and generate printing quality data. The defect matcher adopts a hierarchical structure, including a first matching layer based on the defect state and a second matching layer based on the printing defect type. First, each defect pixel feature is input into the first matching layer, and its geometric feature information, such as contour shape, edge integrity and position offset, is extracted. The geometric deviation between the area and the standard template is calculated by the contour matching algorithm (such as Hu moment feature matching or shape context matching), and the first defect state, such as edge breakage, shape deformation or position offset, is output according to the matching score. Then, according to the inter-layer mapping rule, the defect state generated by the first matching layer is input into the second matching layer, and the grayscale distribution characteristics of the defective pixel area are refined and analyzed in combination with the state. In the second matching layer, the system extracts the grayscale value difference, gradient direction and distribution uniformity, and determines the specific defect type through grayscale histogram matching, correlation calculation or gradient direction deviation analysis. For example, when the grayscale distribution shows local unevenness, it is marked as a stain defect; if the gradient direction is disordered, it is determined as a fuzzy defect; if the grayscale value deviates significantly from the template, it is marked as a color difference defect. After completing the two-layer matching, the system combines the features of each defective pixel with the corresponding defect type to generate detailed record data including defect location, geometric features, grayscale features and defect type, and summarizes all defect data to generate printing quality data.

[0032] Furthermore, the step of constructing a defect matcher includes: Divide the defect states into M categories, where M is a positive integer greater than 1, and the defect states include at least block, line and discrete point; interactively detect the printing detection records, mine the printing defect types, and determine the attribution relationship between the printing defect types and the M categories of defect states, where each category of defect state corresponds to at least one printing defect type; construct a first matching layer based on the M categories of defect states, construct a second matching layer based on the printing defect type, establish an inter-layer mapping between the first matching layer and the second matching layer based on the attribution relationship, and generate a defect matcher.

[0033] When building a defect matcher, the system first classifies the state of printing defects into M types of defect states, where M is a positive integer greater than 1. The defect states include at least block, line and discrete point. Block defects usually manifest as large-area grayscale unevenness, such as stains or missing coatings; Linear defects include broken or blurred printed lines; Discrete point defects refer to scattered abnormal pixels, such as small impurities or color spots. Next, the system loads and interacts with historical printing inspection records, conducts in-depth mining of recorded defect data, and automatically extracts potential printing defect types, such as color difference, stains, blur, offset, etc., through data clustering and pattern analysis algorithms (such as K-means clustering or DBSCAN). At the same time, the system determines the attribution of each printing defect type based on the correspondence between the defect type and the M-type defect state. For example, "stains" belong to block defects, "blurred lines" belong to linear defects, and "impurities" belong to discrete point defects.

[0034] After clarifying the relationship between defect states and printing defect types, the system begins to build a defect matcher. First, the first matching layer is constructed based on M types of defect states. Each state defines a matching model to detect and classify the geometric feature states of the input data. The block defect model focuses on large-area contour features; the linear defect model focuses on line continuity and curvature changes; and the discrete point defect model analyzes the sparse distribution characteristics of pixel points. Subsequently, the system constructs the second matching layer based on the extracted printing defect types. Each defect type is represented by an independent classification model, and the model input includes grayscale value distribution, gradient direction, and texture features. The output of the second matching layer is a specific defect type judgment.

[0035] In order to realize the linkage of two-layer matching, the system establishes inter-layer mapping rules between the first matching layer and the second matching layer based on the aforementioned attribution relationship. The mapping rules stipulate that the matching results of each defect state can trigger a specific defect type classification module. For example, when the first matching layer identifies a "linear defect", the system inputs the corresponding grayscale distribution and gradient features into the second matching layer, further subdividing them into types such as "broken lines" or "blurred lines". Finally, the system integrates the two-layer matching relationship to generate a complete defect matcher for subsequent defect type matching and quality data generation.

[0036] Furthermore, after the printing quality data is output, it includes: According to the periodic printing quality data of a preset period, high-frequency printing defects that meet the preset defect frequency are screened; printing control tracing is performed on the high-frequency printing defects, and a feedback control strategy is determined to guide the printing control of the target printing product.

[0037] After outputting the printing quality data, the system enters the periodic data analysis and feedback control stage. First, the system statistically analyzes the stored printing quality data according to the preset period (such as daily, weekly or monthly), and selects high-frequency printing defects that meet the preset defect frequency. The basis for determining high-frequency printing defects is whether the frequency of occurrence in multiple cycles exceeds the set threshold. For example, the number of occurrences of "stain defects" in a month exceeds 10% of the total number of detections. By counting the frequency of defect types, the system generates a list of high-frequency defects, arranged from high to low according to the frequency of occurrence, highlighting the defects that need to be focused on. For the selected high-frequency printing defects, the system further performs printing control traceability analysis; traceability analysis determines the possible causes by associating the time point of defect occurrence, detection location and related process parameters (such as ink concentration, printing speed, temperature or pressure); the system can call historical process data records and use correlation analysis algorithms (such as Pearson correlation coefficient or multivariate regression analysis) to identify the significant relationship between printing parameters and defect frequency. For example, through analysis, it is found that "color difference defects" are closely related to ink concentration fluctuations, or "line blurring" is related to unstable pressure of the print head.

[0038] After completing the traceability analysis, the system automatically generates feedback control strategies (such as adjusting the upper and lower limits of ink concentration, improving the alignment accuracy of the print head, adding temperature or pressure sensors, etc.) according to the cause of the defect to guide the adjustment of the printing process of the target printed product. For example, the system can automatically send the adjusted ink concentration parameters to the ink control unit of the printing equipment, or trigger the alarm threshold update of the pressure sensor module to ensure that the defect rate is significantly reduced in the subsequent printing process.

[0039] In summary, the embodiments of the present application have at least the following technical effects: First, an image sensor is configured in the printing quality inspection area, and the target printing product is sensed and collected to obtain a printing image. Then, the printing image is transmitted back, and the image is segmented and grayscale processed based on the boundary of the printing product to determine the grayscale printing image. Further, a standard printing template is obtained, and a pixel stack space is constructed to supervise the training of the printing registration module, wherein the printing registration module includes a first pixel registration unit and a second pixel registration unit, and the first acquisition registration is performed by pixel transition sampling, and the second acquisition registration is performed by pixel interval sampling. Then, the grayscale printing image is transmitted to the printing registration module, and the defective pixel features are determined by performing pixel transition sampling registration and pixel interval sampling registration. Finally, the defective pixel features are traversed, and defect type matching is performed based on the defect matcher, and the printing quality data is output. The technical problem of low printing quality detection accuracy and efficiency in the prior art is solved, and the technical effect of improving the printing quality detection accuracy and efficiency is achieved.

[0040] Embodiment 2 is based on the same inventive concept as a high-precision printing quality detection method in the above embodiment. Figure 2 As shown, the present application provides a high-precision printing quality detection system, wherein the system includes: Image acquisition unit 11: configure image sensors in the printing quality inspection area, perform sensing and acquisition on the target printing product to obtain the printing image; image processing unit 12: transmit the printing image back, perform image segmentation and grayscale processing based on the printing product boundary, and determine the grayscale printing image; training unit 13: obtain standard printing templates, and construct pixel stack space, supervise the training of printing registration module, wherein the printing registration module includes a first pixel registration unit and a second pixel registration unit, and performs the first acquisition registration by pixel transition sampling, and performs the second acquisition registration by pixel interval sampling; registration unit 14: transmit the grayscale printing image to the printing registration module, and determine the defective pixel features by performing pixel transition sampling registration and pixel interval sampling registration; defect matching unit 15: traverse the defective pixel features, perform defect type matching based on the defect matcher, and output printing quality data.

[0041] Furthermore, the registration unit 14 is used to perform the following method: The grayscale printed image is transmitted to the first pixel registration unit, and image pixel scanning and pixel transition recognition are performed according to a preset pixel scanning method to locate pixel transition coordinates; according to a preset window scale, window pixel vectors are collected for the pixel transition coordinates, added into the pixel stack space, and a first pixel matrix is ​​determined; according to the standard printing template, a first standard pixel matrix is ​​determined, and the first standard pixel matrix is ​​registered and proofread with the first pixel matrix to locate abnormal pixel areas, wherein the abnormal pixel areas are unregistered window framed areas.

[0042] Furthermore, the registration unit 14 is used to perform the following method: The registration and proofreading uses the pixel position as the first registration target and the grayscale vector as the second registration target, wherein the grayscale vector includes the grayscale value of the pixel transition coordinates and the window grayscale gradient feature, and the window grayscale gradient feature includes the gradient direction and the gradient value; if the abnormal pixel area is empty, a qualified quality result is output; if the abnormal pixel area is not empty, the abnormal pixel area is transferred to the second pixel registration unit.

[0043] Furthermore, the registration unit 14 is used to perform the following method: The abnormal pixel area is transferred to the second pixel registration unit to collect alternate pixel points and add them into the pixel stack space to generate a second pixel matrix; according to the standard printing template, a second standard pixel matrix is ​​determined; the second pixel matrix and the second standard pixel matrix are calibrated to determine pixel defect features, wherein the defective pixel features include geometric features and grayscale distribution features.

[0044] Furthermore, the defect matching unit 15 is used to perform the following method: A defect matcher is constructed, wherein the defect matcher includes a first matching layer based on a defect state and a second matching layer based on a printing defect type; based on the first matching layer, the defective pixel features are identified and the geometric features are matched to determine a first defect state; based on the second matching layer, the grayscale distribution features are matched according to inter-layer mapping to determine the defect type; and the defective pixel features are combined with the defect type to determine the printing quality data.

[0045] Furthermore, the defect matching unit 15 is used to perform the following method: Divide the defect states into M categories, where M is a positive integer greater than 1, and the defect states include at least block, line and discrete point; interactively detect the printing detection records, mine the printing defect types, and determine the attribution relationship between the printing defect types and the M categories of defect states, where each category of defect state corresponds to at least one printing defect type; construct a first matching layer based on the M categories of defect states, construct a second matching layer based on the printing defect type, establish an inter-layer mapping between the first matching layer and the second matching layer based on the attribution relationship, and generate a defect matcher.

[0046] Furthermore, the defect matching unit 15 is used to perform the following method: According to the periodic printing quality data of a preset period, high-frequency printing defects that meet the preset defect frequency are screened; printing control tracing is performed on the high-frequency printing defects, and a feedback control strategy is determined to guide the printing control of the target printing product.

[0047] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0048] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0049] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A high-precision printing quality detection method, characterized in that: The method comprises: An image sensor is configured in the printing quality inspection area to sense and collect the target printing product to obtain the printing image; The printed image is transmitted back, and image segmentation and grayscale processing are performed based on the printed product boundary to determine a grayscale printed image; Obtain a standard printing template, construct a pixel stack space, and supervise the training of a printing registration module, wherein the printing registration module includes a first pixel registration unit and a second pixel registration unit, performs a first acquisition registration by pixel transition sampling, and performs a second acquisition registration by pixel interval sampling; The grayscale printing image is transmitted to the printing registration module, and defective pixel features are determined by performing pixel transition sampling registration and pixel interval sampling registration; The defective pixel features are traversed, defect type matching is performed based on a defect matcher, and printing quality data is output.

2. A high-precision printing quality detection method as claimed in claim 1, characterized in that: The pixel transition sampling and registration comprises: The grayscale printed image is transmitted to the first pixel registration unit, and image pixel scanning and pixel transition recognition are performed according to a preset pixel scanning method to locate pixel transition coordinates; According to a preset window scale, performing window pixel vector acquisition on the pixel transition coordinates, adding the vectors into the pixel stack space, and determining a first pixel matrix; According to the standard printing template, a first standard pixel matrix is ​​determined, and the first standard pixel matrix is ​​aligned and proofread with the first pixel matrix to locate an abnormal pixel area, wherein the abnormal pixel area is an area framed by an unaligned window.

3. A high-precision printing quality detection method as claimed in claim 2, characterized in that: The registration and proofreading uses the pixel position as the first registration target and the grayscale vector as the second registration target, wherein the grayscale vector includes the grayscale value of the pixel transition coordinates and the window grayscale gradient feature, and the window grayscale gradient feature includes the gradient direction and the gradient value; If the abnormal pixel area is empty, output a quality qualified result; If the abnormal pixel region is not empty, the abnormal pixel region is transferred to the second pixel registration unit.

4. A high-precision printing quality detection method as claimed in claim 3, characterized in that: Perform pixel-by-pixel sampling and registration, including: Transferring the abnormal pixel area to the second pixel registration unit, collecting alternate pixel areas and adding them into the pixel stack space to generate a second pixel matrix; Determining a second standard pixel matrix according to the standard printing template; The second pixel matrix and the second standard pixel matrix are checked to determine pixel defect features, wherein the defective pixel features include geometric features and grayscale distribution features.

5. A high-precision printing quality detection method as claimed in claim 4, characterized in that: Defect type matching is performed based on the defect matcher, and printing quality data is output, including: Constructing a defect matcher, the defect matcher comprising a first matching layer based on defect states and a second matching layer based on printing defect types; Based on the first matching layer, identifying the defective pixel feature and matching the geometric feature to determine a first defect state; According to the inter-layer mapping, based on the second matching layer, matching the grayscale distribution characteristics is performed to determine the defect type; The defective pixel feature is combined with the defect type to determine the printing quality data.

6. A high-precision printing quality detection method as claimed in claim 5, characterized in that: The construction of the defect matcher includes: Divide the defect states into M categories, where M is a positive integer greater than 1, and the defect states include at least block, line and discrete point; Interactive printing inspection records, mining printing defect types, and determining the attribution relationship between the printing defect types and M types of defect states, wherein each type of defect state corresponds to at least one printing defect type; A first matching layer is constructed based on the M-type defect state, a second matching layer is constructed based on the printing defect type, an inter-layer mapping between the first matching layer and the second matching layer is established based on the attribution relationship, and a defect matcher is generated.

7. A high-precision printing quality detection method as claimed in claim 1, characterized in that: After outputting the printing quality data, it includes: According to the periodic printing quality data of the preset period, high-frequency printing defects that meet the preset defect frequency are screened; Printing control source tracing is performed for the high-frequency printing defects, a feedback control strategy is determined, and printing control of the target printing product is guided.

8. A high-precision printing quality inspection system, characterized in that: A system for implementing a high-precision printing quality detection method according to any one of claims 1 to 7, the system comprising: Image acquisition unit: an image sensor is configured in the printing quality inspection area to perform sensing and acquisition on the target printing product to obtain the printing image; Image processing unit: transmits the printed image back, performs image segmentation and grayscale processing based on the printed product boundary, and determines the grayscale printed image; Training unit: obtaining a standard printing template, constructing a pixel stack space, and supervising the training of a printing registration module, wherein the printing registration module comprises a first pixel registration unit and a second pixel registration unit, performing a first acquisition registration by pixel transition sampling, and performing a second acquisition registration by pixel interval sampling; Registration unit: transmitting the grayscale printing image to the printing registration module, and determining defective pixel features by performing pixel transition sampling registration and pixel interval sampling registration; Defect matching unit: traverses the defect pixel features, performs defect type matching based on the defect matcher, and outputs printing quality data.

Images