Detection method of large printing plate

Abstract

One embodiment of the method comprises the following steps: collecting a visual image or a target electronic image of a large printing edition, matching according to a preset page shape, page direction, page size and page logic position, and determining a detection area in the visual image or the target electronic image; segmenting the detection area through an image semantic segmentation model to obtain a character area, a picture area, a color block area and a blank area; analyzing the character area, the picture area, the color block area and the blank area according to respective analysis methods of the character area, the picture area, the color block area and the blank area; and outputting one or more of actual positions, contours, directions and sizes of the character region, the picture region, the color block region and the blank region in the printing large plate based on the analysis result.

Description

Technical Field

[0001] The present disclosure relates to the field of printing automation technologies, and in particular, to a method, apparatus, system, storage medium, and computer program product for detecting a large-format printing file image. Background Art

[0002] Large-format printing refers to a process of using a large-format printing press (usually with a format exceeding A1 size, such as folio, full sheet, or larger) to print multiple pages of content or large-format printed products at one time. By combining multiple pages into a large format, which is called "page imposition", the production efficiency can be significantly improved and the cost can be reduced. The core processes of large-format printing include: page imposition, that is, using professional software (such as Preps, Signastation) to arrange multiple pages onto the large format according to the folding and binding methods; plate making (CTP technology), that is, outputting digital files to printing plates through a computer-to-plate (CTP) machine; printing control and inspection, that is, ensuring the accuracy of ink presetting, multi-color printing alignment, and inspecting the printing quality through software management. Currently, the above core processes have all achieved a high degree of automation. During the automation process, it is necessary to extract the actual physical information of the content to be printed in the large-format printing, such as: the actual position, contour, orientation, size, etc. of the area to be printed in the large-format printing.

[0003] Especially in the page imposition process, multiple small printing pages or printing single modules are assembled into a large-format printing electronic file in a pre-press automation system. It is necessary to quickly and accurately map the small printing pages or printing single modules to the large-format printing to meet the automation requirements. There are certain deformations between the printed product image and the electronic large-format image, resulting in the inability of the single-point alignment method to eliminate the objective deformation of the paper and unable to better perform precise positioning and matching between the electronic file and the large-format scanned image. Currently, when the electronic file image provided by the electronic file producer or provider does not carry information such as its logical position, physical page number sequence, contour, orientation, size, etc. in the large-format printing file image, the small printing page or printing single module file and image are matched using the image similarity method. The success rate of the image similarity matching cannot reach 100%, and it cannot ensure successful matching every time. Moreover, the matching speed is very slow and cannot meet the requirements of automated processing.

[0004] An existing technical solution is to read the standard file images of the imposition and folding marks of the reading system (for example: XML files, .jdf, .tpl, .tplx format files that comply with the regulations of the JDF organization, and other related descriptive format files, etc.), and pre-position the information such as the logical position, physical page number sequence, outline, orientation, and size of the printed small page or printed single mode in the printed large format. That is, the provider of the electronic file image or the processor of the electronic file image needs to extract the relevant information such as the logical position, physical page number sequence, outline, orientation, and size of the printed small page or printed single mode file in a single printed large format and a batch of printed large formats from the imposition and folding mark template file, and obtain the CTM transformation matrix of the relevant information to automatically realize the automatic matching and mutual mapping of the printed small page or printed single mode file and the corresponding printed large format file and image. However, this method only assigns the mapping relationship of the entire printed small page or printed single mode in the printed large format. In the later processing, only the standard file image of the entire printed small page or printed single mode can be used to perform an overall image comparison with the image of the corresponding area of the printed large format. In the pixel-level processing and comparison, the accuracy and processing speed of this method still cannot meet the expectations.

[0005] The above problems are particularly prominent in the quality inspection of printed images. In the printing process flow, defects such as file errors, modification errors, file pixel loss, incorrect loading of printed small pages in imposition, layer pixel loss, and graphic interpretation errors during rasterization often occur. Therefore, in each printing process, there is a need for comparative quality inspection between the original file and the next form of the original file to ensure that the expected graphic content is correctly transferred step by step onto the printing substrate such as paper. Therefore, the insufficient accuracy and processing speed of image processing seriously affect the quality of printed image quality inspection.

[0006] The above information is only presented as background information to help understand the present disclosure. No confirmation or other relevant indication has been made as to whether any of the above can be used as the prior art for the present disclosure. Summary of the Invention

[0007] Embodiments of the present disclosure solve the problems in the detection of the printed large format file image described above. A method for detecting a printed large format file image with high response speed, high accuracy, and high reliability is provided. A device for detecting a printed large format file image, a system for detecting a printed large format file image, a non-transitory storage medium, and a computer program product are provided to solve the problems in the detection of the printed large format file image.

[0008] The first aspect of the present disclosure provides a method for detecting a printed large - format document image, including: collecting a visual image or a target electronic image of the printed large - format, performing matching according to a preset page shape, page orientation, page size, and page logical position to determine a detection area in the visual image or the target electronic image; segmenting the detection area through an image semantic segmentation model to obtain a text area, a picture area, a color block area, and a blank area; analyzing the text area, the picture area, the color block area, and the blank area according to their respective analysis methods; and based on the results of the analysis, outputting one or more of the actual positions, contours, orientations, and sizes of the text area, the picture area, the color block area, and the blank area in the printed large - format.

[0009] For example, in at least one embodiment, before segmenting the detection area through the image semantic segmentation model, multiple actual detection areas are labeled to obtain multiple labeled maps, and the multiple labeled maps include the contours and labeling features of the text area, the picture area, the color block area, and the blank area; using the multiple labeled maps as training samples and the labeling features as attribute labels, training the image semantic segmentation model to recognize the text area, the picture area, the color block area, and the blank area for different regions.

[0010] For example, in at least one embodiment, the preset page shape, page orientation, page size, and page logical position are generated based on a standard document image from which the printed large - format is derived; the preset page shape, page orientation, page size, and page logical position have a mapping relationship with the standard document image.

[0011] For example, in at least one embodiment, the standard document image is segmented in advance by the image semantic segmentation model to obtain the text region, the picture region, the color block region, and the blank region in the standard document image; segmenting the detection region by the image semantic segmentation model includes: directly segmenting the detection region in the visual image or the target electronic image by the image semantic segmentation model, or, after segmenting the standard document image based on the image semantic segmentation model, obtaining the text region, the picture region, the color block region, and the blank region and the mapping relationship in the standard document image, and mapping the text region, the picture region, the color block region, and the blank region in the detection region in the visual image or the target electronic image; based on different comparison strategies, respectively comparing the detection region with the text region, the picture region, the color block region, and the blank region of the standard document image to determine the comparison results of their respective regions; and transmitting the comparison results of each region to the output device.

[0012] For example, in at least one embodiment, the printing large plate is composed of multiple printing small pages or printing single modes; the matching further includes: obtaining the detection region according to the preset page logical position of the printing small page or the printing single mode in the printing large plate; the preset page position, page shape, page direction, and page size are the respective corresponding page position, page shape, page direction, and page size generated by each printing small page or printing single mode based on the standard document image; the detection region is the respective corresponding detection region of each printing small page or printing single mode.

[0013] For example, in at least one embodiment, matching according to the preset page shape, page direction, page size, and page logical position to determine the detection region in the visual image or the target electronic image includes: establishing a mapping relationship between the printing small page or the printing single mode and the corresponding region in the image of the printing large plate according to the imposition rules, where the mapping relationship includes mapping the number information of the printing small page or the printing single mode to the page position, page shape, page direction, and page size in the printing large plate.

[0014] For example, in at least one embodiment, the analysis method of the text region includes: performing a primary analysis on one or more of the actual position, contour, direction, and size of the text region; correcting the visual image of the text region obtained by the primary analysis according to the text correction parameter; and performing a secondary analysis on one or more of the actual position, contour, direction, and size of the text region in the corrected visual image.

[0015] For example, in at least one embodiment, the analysis method of the picture area includes: performing a first analysis on one or more of the actual position, contour, orientation, and size of the picture area; correcting the visual image of the text area obtained from the first analysis according to picture correction parameters; and performing a second analysis on one or more of the actual position, contour, orientation, and size of the picture area in the corrected visual image.

[0016] For example, in at least one embodiment, the analysis method of the text area and the picture area includes: performing a first analysis on one or more of the actual position, contour, orientation, and size of the text area and the picture area; calculating correction parameters according to the results of the first analysis, and performing a second correction on the correction parameters to obtain text correction parameters and picture correction parameters respectively, where the picture correction parameters are greater than the text correction parameters; correcting the visual images of the text area and the picture area obtained from the first analysis according to the text correction parameters and the picture correction parameters; and performing a second analysis on one or more of the actual position, contour, orientation, and size of the text area and the picture area in the corrected visual images.

[0017] A second aspect of the present disclosure provides a detection device for a printed large-format document image, including: an acquisition module configured to acquire a visual image or a target electronic image of a printed large-format, match according to preset page shape, page orientation, page size, and page logical position, and determine a detection area in the visual image or the target electronic image; a segmentation module configured to segment the detection area through an image semantic segmentation model to obtain a text area, a picture area, a color block area, and a blank area; an analysis module configured to analyze the text area, the picture area, the color block area, and the blank area according to the respective analysis methods of the text area, the picture area, the color block area, and the blank area; and an output module configured to output one or more of the actual position, contour, orientation, and size of the text area, the picture area, the color block area, and the blank area in the printed large-format based on the detection results.

[0018] A third aspect of the present disclosure provides a detection system for a printed large-format document image, including: a memory for non-temporarily storing computer-executable instructions; and a processor for running the computer-executable instructions, where the computer-executable instructions, when run by the processor, execute the detection method for a printed large-format document image according to any one of the above.

[0019] A fourth aspect of the present disclosure provides a non - transitory storage medium that non - transitorily stores computer - executable instructions. When the computer - executable instructions are executed by a computer, the method for detecting a printed large - format file image according to any one of the above is executed.

[0020] A fifth aspect of the present disclosure is a computer program product, including a computer program, characterized in that when the computer program is executed by a processor, the method for detecting a printed large - format file image according to any one of the above is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the accompanying drawings of the embodiments will be briefly introduced below. Obviously, the drawings in the following description only relate to some embodiments of the present disclosure and do not limit the present disclosure.

[0022] Figure 1 is a flowchart of detecting a printed large - format file image according to an embodiment of the present disclosure;

[0023] Figure 2 is a flowchart of detecting a second printed large - format file image according to an embodiment of the present disclosure;

[0024] Figure 3 is a schematic diagram of dividing a file image area according to an embodiment of the present disclosure;

[0025] Figure 4 is a flowchart of detecting a third printed large - format file image according to an embodiment of the present disclosure;

[0026] Figure 5 is a flowchart of detecting a fourth printed large - format file image according to an embodiment of the present disclosure;

[0027] Figure 6 is a schematic diagram of the mapping relationship between a printed small page and a printed large - format measurement area according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present disclosure clearer, the technical solutions of the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present disclosure. Obviously, the described embodiments are some, but not all, of the embodiments of the present disclosure. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the described embodiments of the present disclosure fall within the scope of protection of the present disclosure.

[0029] Unless otherwise defined, technical terms or scientific terms used in this disclosure shall have the ordinary meanings understood by those of ordinary skill in the art to which this disclosure pertains. The terms "first", "second" and similar terms used in this disclosure do not denote any order, quantity or importance, but are only used to distinguish different components. Words such as "comprising" or "including" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. Words such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. Words such as "upper", "lower", "left" and "right" are only used to indicate relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly. In this disclosure, "a plurality of" means two or more.

[0030] According to an embodiment of the present disclosure, Figure 1 a method for detecting a printed large-format document image is disclosed, comprising the following steps:

[0031] S101: Collect a visual image or a target electronic image of the printed large-format. Among them, the visual image refers to an image that can be directly recognized by the human eye or a visual device, which is visually displayed through a display device in computer software or on a printing machine. The target electronic image refers to an electronic file stored in a storage medium as a detection object, which includes the image information of the target to be detected. The visual image or the target electronic image can be an image output from any one of the processes such as printed large-format document editing, imposition, halftone rasterization, printing, etc., and is derived from an electronic file or a physical printed matter.

[0032] Preferably, the visual image or the target electronic image of the printed large-format can be collected through imposition analysis software. By connecting the imposition analysis software, the visual image or the target electronic image of the printed large-format can be directly obtained. Directly obtaining the visual image or the target electronic image of the printed large-format through connecting the imposition analysis software is beneficial to directly applying the detection result to the result optimization and quality inspection of the connected imposition analysis software, and simplifies the difficulty of transcoding and other work between the systems.

[0033] Preferably, the visual image or the target electronic image of the printed large-format can be collected through image recognition software. By the image recognition software, the image information of the printed large-format can be directly input; or through an external collection device, such as a scanner, a camera, etc., the image information of the physical printed matter of the printed large-format can be recognized. Collecting through the image recognition software can simply and directly obtain visual information.

[0034] Preferably, the large-format printed image can be a halftone image in a 1-bit-tiff file. The halftone image is interpolated by the imposition analysis software to restore and generate a continuous-tone image. Preferably, when generating the continuous-tone image, the pixel positions of the continuous-tone image are linearly mapped to the halftone image, and then bicubic interpolation is performed according to the neighboring pixels of the halftone image to generate the continuous-tone image. The bicubic interpolation coefficient is calculated based on the mapped image position and the positions of the neighboring pixels.

[0035] S102: Match according to the preset page shape, page orientation, page size, and page logical position to determine the detection area in the visual image or the target electronic image. Preferably, the page shape, page orientation, and page size are the actual shape, orientation, and size of the page in the large-format print. The page logical position information includes the serial number of the large-format print corresponding to the page among multiple large-format prints, and the actual position of the page in the large-format print with the corresponding serial number.

[0036] Preferably, as Figure 2 shown, the following steps are also included in step S102:

[0037] S1021: Generate the preset page shape, page orientation, page size, and page logical position based on the standard file image sourced from the large-format print.

[0038] Preferably, the standard file image is the final image that is desired to be printed before plate making and is the standard image of the printed matter.

[0039] Preferably, the standard file image is an electronic file containing image information obtained before any process before acquiring the visual image or the target electronic image, or a physical image of the printed matter obtained by means such as scanning or photographing.

[0040] S1022: Match according to the page shape, page orientation, page size, and page logical position sourced from the standard file image to determine the detection area in the visual image or the target electronic image. Preferably, obtain the matching relationship between the standard file image and the corresponding detection area, and use this matching relationship to align the standard file image with the corresponding detection area.

[0041] Preferably, the pixel value R of the corresponding detection area is obtained, and the number of matching feature points between the standard document image and the corresponding detection area is determined according to the comparison result between the pixel value R and the preset pixel value. There are a first preset pixel value R1, a second preset pixel value R2, a first number of feature points W1, a second number of feature points W2, and a third number of feature points W3, where R1 < R2 and W1 < W2 < W3. When R ≤ R1, the image matching unit determines that the number of matching feature points is W1; when R1 < R ≤ R2, the image matching unit determines that the number of matching feature points is W2; when R > R2, the image matching unit determines that the number of matching feature points is W3. By setting multiple preset pixel values and the number of feature points, and determining the number of matching feature points corresponding to the matching between the standard document image and the corresponding detection area according to the comparison result between the pixel value of the corresponding detection area and the preset pixel value, this alignment method improves the data processing accuracy during intelligent comparison, thereby further improving the quality of printed matter.

[0042] S1023: A mapping relationship is formed between the standard document image and the detection area. Through the above steps, preset information is formed based on the standard document image, and then accurate information of the corresponding detection area is generated based on the preset information. At this time, an accurate mapping relationship is formed between the standard document image and the detection area. Based on the mapping relationship, in the standard document image, such as various vector and image format files like PDF, tiff, png, jpg, etc., a CTM transformation matrix with relevant information such as the printing plate code, page shape, page orientation, page size, and page logical position of the corresponding detection area in the printing plate is added. The standard document image with the CTM transformation matrix added can be used as the basic information for subsequent processes such as segmentation, further image processing, especially for processes with high precision requirements such as image quality inspection, imposition, and mapping graphic and text information.

[0043] As Figure 1 shown, after step S102, it further includes S103: The detection area is segmented by an image semantic segmentation model to obtain a text area, a picture area, a color block area, and a blank area. As Figure 3 shown, usually, the pattern on a printed matter includes a text area, a picture area, a color block area, and a blank area. In the field of computer image comparison, the optimal processing strategies for different areas are obviously different. Therefore, dividing different areas, aligning them separately, and then processing them is beneficial to improving the efficiency and accuracy of image processing.

[0044] Preferably, a semantic segmentation model is used to segment the detection area. Among them, semantic segmentation aims to assign a semantic class label (such as "person", "car", "background", etc.) to each pixel in the image. Different from instance segmentation (distinguishing different individuals of the same kind of object), semantic segmentation usually only focuses on class discrimination at the pixel level. Semantic segmentation models in the prior art, such as PSPNet (Pyramid Scene Parsing Network), DeepLab series (v1 - v3+), U - Net, SegNet, etc., can be used to segment the detection area.

[0045] Preferably, as Figure 4 shown, in step S103, using a semantic segmentation model to segment the detection area further includes the following steps:

[0046] S1031: Label multiple actual detection areas to obtain multiple labeled maps, which include the outlines and labeling features of text areas, picture areas, color block areas, and blank areas. Among them, multiple actual detection areas refer to images on multiple printing plates that can be used as actual detection areas, and these images together constitute the data set for training the semantic segmentation model. Normalize or standardize the images in the data set, and further enhance the processed data to obtain multiple labeled maps. Convert the labeled maps into single - channel class index maps to handle possible class imbalance problems.

[0047] S1032: Use multiple labeled maps as training samples and the labeling features as attribute labels to train the image semantic segmentation model to recognize text areas, picture areas, color block areas, and blank areas for different regions. Select a semantic segmentation model such as one of the aforementioned ones, select appropriate loss functions and evaluation metrics, and train the semantic segmentation model. Preferably, continuously optimize and debug during the training of the language training model to improve the efficiency and accuracy of the model.

[0048] Preferably, the detection area can be segmented directly using a semantic segmentation model or indirectly using other methods.

[0049] Preferably, it includes S10331: Use the trained semantic segmentation model to directly segment the already determined detection area in the printing plate image.

[0050] Preferably, it can also include S10332: Pre - segment the standard document image through the image semantic segmentation model to obtain text areas, picture areas, color block areas, and blank areas in the standard document image.

[0051] After step S10332, it includes S10342: After segmenting the standard document image based on the image semantic segmentation model, the text regions, picture regions, color block regions, and blank regions and their mapping relationships obtained in the standard document image are mapped to the text regions, picture regions, color block regions, and blank regions in the detection region of the visual image or the target electronic image. In the previous steps, an accurate mapping relationship has been generated between the standard document image and the printed image of the printing plate, that is, the visual image or the target electronic image. Based on this mapping relationship, the segmentation result of the standard document image can be mapped to the visual image or the target electronic image.

[0052] As Figure 1 shown, after step S103, it further includes S104: According to the respective analysis methods of the text regions, picture regions, color block regions, and blank regions, the text regions, picture regions, color block regions, and blank regions are analyzed. Based on the semantic segmentation result, for the concerns of different regions in different process requirements, the elements or parameters of the downstream process requirements are analyzed, such as one or more of the actual positions, contours, directions, and sizes of each region in the printing plate. The specific analysis methods adopted for each region can be the same or different.

[0053] Preferably, different analysis methods are used for the text regions, picture regions, color block regions, and blank regions.

[0054] When inspecting the image collected after the printing plate is actually printed by the printing press, the actual situations of the images in different regions are much more complex. Usually, the necessary fluidity of the printing ink must be maintained. If the fluidity of the ink is insufficient, the condition of the intermediate break of the ink layer will be damaged, resulting in insufficient remaining ink on the plate, causing the defect of the printing plate due to the exposure and wear of the graphic part. Therefore, in order to avoid the defect of the printing plate and to maintain the necessary fluidity of the ink, and since the ink has a certain degree of rendering property, excessive ink will inevitably accumulate at some glyphs, resulting in thickening and fuzzing of the font. The quality of the printing material will also affect the printing effect. The printing material cannot achieve absolute surface flatness and uniform thickness, which will also cause thickening and fuzzing of some fonts. Different paper types, different viscosities of the ink, and different printing speeds will all result in different degrees of thickening and fuzzing. In printed products, under the condition of maintaining appropriate ink fluidity, the rendering of the picture region is usually stronger than that of the text region.

[0055] Preferably, the analysis method for the text area includes: performing a first analysis on one or more of the actual position, contour, orientation, and size of the text area. Correcting the visual image of the text area obtained from the first analysis according to the text correction parameters. In the corrected visual image, performing a second analysis on one or more of the actual position, contour, orientation, and size of the text area. After the first analysis, the analysis software can calculate the degree of thickening and fuzzing of the text in the printed matter, and obtain appropriate text correction parameters through fitting of the entire area. In pixel-level image analysis, performing a second analysis on the text area based on the text correction parameters is conducive to more accurate alignment of the standard document image and the measured area, and can effectively reduce misjudgment and noise in subsequent processing.

[0056] Preferably, the analysis method for the picture area includes: performing a first analysis on one or more of the actual position, contour, orientation, and size of the picture area. Correcting the visual image of the text area obtained from the first analysis according to the picture correction parameters. In the corrected visual image, performing a second analysis on one or more of the actual position, contour, orientation, and size of the picture area. After the first analysis, the analysis software can calculate the degree of thickening and fuzzing of the picture in the printed matter, and obtain appropriate picture correction parameters through fitting of the entire area. In pixel-level image analysis, performing a second analysis on the picture area based on the picture correction parameters is conducive to more accurate alignment of the standard document image and the measured area, and can effectively reduce misjudgment and noise in subsequent processing.

[0057] Preferably, the analysis method for the text area and the picture area includes: performing an analysis on one or more of the actual position, contour, orientation, and size of the text area and the picture area. According to the result of the first analysis, correction parameters are calculated, and the correction parameters are secondarily corrected to obtain a text correction parameter and a picture correction parameter respectively, where the picture correction parameter is greater than the text correction parameter. According to the text correction parameter and the picture correction parameter, the visual images of the text area and the picture area obtained from the first analysis are corrected. In the corrected visual image, a secondary analysis is performed on one or more of the actual position, contour, orientation, and size of the text area and the picture area. After the first analysis, the analysis software can calculate the degree of thickening and fuzzing of the text and pictures in the printed matter, and through the fitting of the entire area, appropriate correction parameters are obtained. According to the different degrees of thickening and fuzzing of the text area and the picture area of different printed matters, the correction parameters are secondarily corrected to obtain different text correction parameters and picture correction parameters. The picture correction parameter being greater than the text correction parameter is conducive to more accurately determining the characteristics of the picture area and the text area, especially more accurately determining the contour lines of the two areas. In pixel-level image analysis, performing a secondary analysis on the text area and the picture area based on the text correction parameter and the picture correction parameter is conducive to more precise alignment of the standard document image and the measured area, and can more effectively reduce misjudgment and noise in subsequent processing.

[0058] As Figure 1 shown, after step S104, it further includes S105: Based on the analysis result, output one or more of the actual position, contour, orientation, and size of the text area, the picture area, the color block area, and the blank area in the printing plate. The analysis result is the result to be output, and it can be output to the receiving devices required by downstream processes, such as: imposition software, imposition machine, display device, etc.

[0059] The above detection method can achieve pixel-level accuracy, and the detected position of the detection area is accurate, which can effectively solve problems such as misalignment, and is widely used in fields such as image quality inspection, imposition, and mapping graphic and text information, which is conducive to improving printing quality, increasing printing accuracy, reducing defects, and improving the automation level and processing speed of corresponding processes.

[0060] A typical implementation method is, for example, in the field of image quality inspection, as Figure 5 shown, in parallel with steps S101 - S105, it further includes:

[0061] S201: Segment the standard document image using an image semantic segmentation model to obtain text regions, picture regions, color block regions, and blank regions in the standard document image. The actual content of step S201 is the same as the possible sub-step S10332 in the aforementioned step S103. If both steps appear in the same embodiment, only one step may be executed, and the recorded results may be used for the other step.

[0062] S202: Analyze the text regions, picture regions, color block regions, and blank regions according to their respective analysis methods in the standard document image. Analyze the standard document image to obtain the elements and parameters that need attention. In the field of image quality inspection, the positions, contours, directions, and sizes of different regions need to be accurate to the pixel level.

[0063] S203: Based on the analysis results, output one or more of the actual positions, contours, directions, and sizes of the text regions, picture regions, color block regions, and blank regions in the printing plate. Output the image detection results for the standard document image.

[0064] After steps S105 and S203, there is also step 106: Based on different comparison strategies, compare the detected regions with the text regions, picture regions, color block regions, and blank regions of the standard document image respectively to determine the comparison results of their respective regions. Based on the characteristics of printed matter, different regions focus on different comparison strategies. For example, for text regions, compare whether there are errors such as similar-shaped characters and adjacent strokes; for picture regions, whether there are defects such as incorrect color tones and uneven lines; for color block regions, whether there are incorrect color tones, extra patterns, ink residues, etc.; for blank regions, whether there are extra patterns, ink residues, etc. Comparing based on different comparison strategies is conducive to improving the comparison efficiency, improving the comparison accuracy, and greatly reducing misjudgments.

[0065] After step S106, there is also step 107: Transmit the comparison results of each region to the output device. For example, output the comparison results to a display device to highlight the quality inspection results. So as to correct defects and eliminate misjudgments.

[0066] Preferably, as Figure 6As shown, the large printing plate is composed of multiple small printing pages or printing single modules. The matching also includes: obtaining a detection area according to the preset page logical position of the small printing page or printing single module in the large printing plate. The preset page position, page shape, page direction, and page size are the respective corresponding page position, page shape, page direction, and page size of each small printing page or printing single module generated based on the standard document image. The detection area is the respective corresponding detection area of each small printing page or printing single module. Usually, the large printing plate is composed of small printing pages. Since the printing press has a large format, multiple small printing pages need to be combined onto a large plate that meets the printing press format according to specific rules to improve printing efficiency and control costs. In the plate-making process, the page number sequence needs to be determined according to the folding method, and a folding sample needs to be made to verify the layout logic. Figure 6 An exemplary folding method is listed. At the same time, a bleed edge needs to be left at the splicing position of single-page printed matter to ensure that there is no white edge after cutting. Figure 6 It can be seen that the physical characteristics of the small printing pages in the large printing plate are different from each other. Using the detection method of the embodiments of the present disclosure is beneficial to the accurate matching and detection of each small printing page in the large printing plate.

[0067] Preferably, the matching is performed according to the preset page shape, page direction, page size, and page logical position to determine the detection area in the visual image or target electronic image, including: establishing a mapping relationship between the small printing page or printing single module and the corresponding area in the image of the large printing plate according to the plate-making rules. The mapping relationship includes mapping the number information of the small printing page or printing single module to the page position, page shape, page direction, and page size in the large printing plate. Based on the characteristics that the small printing pages form the large printing plate through the plate-making process, according to the standard document image of the small printing page, a mapping relationship between the standard document image and the corresponding detection area in the large printing plate is formed, and the number corresponding to the large printing plate, the page shape, page direction, page size, and page logical position of the detection area in the large printing plate are accurately determined at the pixel level, which is beneficial to improving the accuracy and efficiency of subsequent processes such as plate-making and image inspection.

[0068] According to an embodiment of the present disclosure, there is also provided a detection device for a printed large-format file image, including: a collection module configured to collect a visual image or a target electronic image of the printed large-format. A determination module configured to determine a detection area in the visual image or the target electronic image by matching according to a preset page shape, page orientation, page size, and page logical position. A segmentation module configured to segment the detection area through an image semantic segmentation model to obtain a text area, a picture area, a color block area, and a blank area. An analysis module configured to analyze the text area, the picture area, the color block area, and the blank area according to their respective analysis methods. An output module configured to output one or more of the actual positions, contours, orientations, and sizes of the text area, the picture area, the color block area, and the blank area in the printed large-format based on the detection results.

[0069] According to an embodiment of the present disclosure, there is also provided a detection system for a printed large-format file image, including: a memory for non-temporarily storing computer-executable instructions; and a processor for running the computer-executable instructions, wherein when the computer-executable instructions are run by the processor, the detection method for the printed large-format file image of any one of the above embodiments is executed.

[0070] According to an embodiment of the present disclosure, there is also provided a non-temporary storage medium that non-temporarily stores computer-executable instructions, wherein when the computer-executable instructions are executed by a computer, the detection method for the printed large-format file image of any one of the above embodiments is executed.

[0071] According to an embodiment of the present disclosure, there is also provided a computer program product, including a computer program, characterized in that when the computer program is executed by a processor, the detection method for the printed large-format file image of any one of the above embodiments is implemented.

[0072] There are also the following points to note:

[0073] (1) The drawings of the embodiments of the present disclosure only relate to the structures involved in the embodiments of the present disclosure, and other structures can refer to the general design.

[0074] (2) For clarity, in the drawings used to describe the embodiments of the present disclosure, the thicknesses of devices, layers, or regions are enlarged or reduced, that is, these drawings are not drawn according to the actual scale. It can be understood that when an element such as a layer, a film, a region, or a substrate is referred to as being "on" or "under" another element, the element can be "directly" on or under the other element or there can be intermediate elements.

[0075] (3) Without conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other to obtain new embodiments.

[0076] The above are only specific embodiments of the present disclosure, but the protection scope of the present disclosure is not limited thereto. The protection scope of the present disclosure shall be subject to the protection scope of the claims.

Claims

1. A detection method for printing large - format document images, comprising: Collecting visual images or target electronic images of the printing large - format; Matching according to preset page shape, page orientation, page size, and page logical position to determine the detection area in the visual image or the target electronic image; Segmenting the detection area through an image semantic segmentation model to obtain text areas, picture areas, color block areas, and blank areas; Analyzing the text areas, picture areas, color block areas, and blank areas according to their respective analysis methods; Based on the results of the analysis, outputting one or more of the actual positions, contours, orientations, and sizes of the text areas, picture areas, color block areas, and blank areas in the printing large - format.

2. The detection method of the printing large-format document image according to claim 1, wherein, Before segmenting the detection area through the image semantic segmentation model, multiple actual detection areas are labeled to obtain multiple labeled maps, and the multiple labeled maps include the contours and labeling features of the text areas, picture areas, color block areas, and blank areas; Using the multiple labeled maps as training samples and the labeling features as attribute labels, training the image semantic segmentation model to recognize the text areas, picture areas, color block areas, and blank areas for different regions.

3. The detection method of the printed large-format document image according to claim 2, wherein, The preset page shape, page orientation, page size, and page logical position are generated based on the standard document image from which the printing large - format is derived; The preset page shape, page orientation, page size, and page logical position have a mapping relationship with the standard document image.

4. The detection method of the printing large-format document image according to claim 3, wherein, Pre - segmenting the standard document image through the image semantic segmentation model to obtain the text areas, picture areas, color block areas, and blank areas in the standard document image; Segmenting the detection area through the image semantic segmentation model includes: directly segmenting the detection area in the visual image or the target electronic image through the image semantic segmentation model, Or, after segmenting the standard document image based on the image semantic segmentation model, mapping the text areas, picture areas, color block areas, and blank areas obtained in the standard document image and the mapping relationship to map out the text areas, picture areas, color block areas, and blank areas in the detection area of the visual image or the target electronic image; Based on different comparison strategies, respectively comparing the detection area with the text areas, picture areas, color block areas, and blank areas of the standard document image to determine the comparison results of each area; Transmitting the comparison results of each area to an output device.

5. The detection method of the printed large-format document image according to any one of claims 1 to 4, wherein, The printing large - format is composed of multiple printing small pages or printing single modules; The matching further includes: obtaining the detection area according to the preset page logical position of the printing small page or the printing single module in the printing large - format. The preset page position, page shape, page orientation, and page size are the corresponding page position, page shape, page orientation, and page size of each of the printing sub-pages or the printing single mode generated based on the standard document image; The detection area is the corresponding detection area of each of the printing sub-pages or the printing single mode.

6. The detection method of the printing large-format file image according to claim 5, wherein, Matching according to the preset page shape, page orientation, page size, and page logical position to determine the detection area in the visual image or the target electronic image, including: According to the imposition rules, establish a mapping relationship between the corresponding areas in the image of the printing sub-page or the printing single mode and the printing large format. The mapping relationship includes mapping the number information of the printing sub-page or the printing single mode to the page position, page shape, page orientation, and page size in the printing large format.

7. The method for detecting a printing large format document image according to claim 1, Among them, The analysis method of the text area includes: performing a single analysis on one or more of the actual position, contour, orientation, and size of the text area; According to the text correction parameter, correct the visual image of the text area obtained from the single analysis; In the corrected visual image, perform a secondary analysis on one or more of the actual position, contour, orientation, and size of the text area.

8. The method for detecting a printing large format document image according to claim 1, Among them, The analysis method of the picture area includes: performing a single analysis on one or more of the actual position, contour, orientation, and size of the picture area; According to the picture correction parameter, correct the visual image of the text area obtained from the single analysis; In the corrected visual image, perform a secondary analysis on one or more of the actual position, contour, orientation, and size of the picture area.

9. The method for detecting a printing large format document image according to claim 1, Among them, The analysis method of the text area and the picture area includes: performing a single analysis on one or more of the actual position, contour, orientation, and size of the text area and the picture area; According to the result of the single analysis, calculate the correction parameter, perform a secondary correction on the correction parameter, and respectively obtain the text correction parameter and the picture correction parameter, where the picture correction parameter is greater than the text correction parameter; According to the text correction parameter and the picture correction parameter, correct the visual images of the text area and the picture area obtained from the single analysis; In the corrected visual image, perform a secondary analysis on one or more of the actual position, contour, orientation, and size of the text area and the picture area.

10. A device for detecting a printing large format document image, comprising: An acquisition module configured to acquire a visual image or a target electronic image of the printing large format; A determination module, configured to determine a detection area in the visual image or the target electronic image by matching according to a preset page shape, page orientation, page size, and page logical position; A segmentation module, configured to segment the detection area through an image semantic segmentation model to obtain a text area, a picture area, a color block area, and a blank area; An analysis module, configured to analyze the text area, the picture area, the color block area, and the blank area according to the respective analysis methods of the text area, the picture area, the color block area, and the blank area; An output module, configured to output one or more of the actual positions, contours, orientations, and sizes of the text area, the picture area, the color block area, and the blank area in the printing large format based on the detection results.

11. A detection system for a printing large format file image, comprising: A memory, configured to non-temporarily store computer-executable instructions; And a processor, configured to run the computer-executable instructions, wherein the computer-executable instructions, when run by the processor, execute the detection method for a printing large format file image according to any one of claims 1-9.

12. A non-transitory storage medium that non-transitorily stores computer-executable instructions, wherein, When the computer-executable instructions are executed by a computer, the detection method for a printing large format file image according to any one of claims 1-9 is executed.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, the detection method for a printing large format file image according to any one of claims 1-9 is implemented.

Images