Printed matter image monitoring method based on optical recognition technology

Abstract

The present application relates to the technical field of image processing and pattern recognition, and particularly relates to a printed matter image monitoring method based on optical recognition technology, which comprises the following steps: collecting visible light, near-infrared, ultraviolet fluorescence and grazing light images of printed matter, and synthesizing into a standardized multi-modal data set; calculating paper structure density index and ink chemical stability index respectively through region division, and generating a spatial distribution map; calculating a change field and a characteristic degradation rate based on a time series map, and dividing four risk levels by comprehensively considering current state and change trend; predicting residual life based on a first-order reaction kinetics equation and the characteristic degradation rate; and finally generating a health diagnosis comprehensive view which integrates state, change, risk and life information. The present application converts optical signals into quantifiable material health indicators, realizes a technical leap from static recognition to dynamic diagnosis to trend prediction, and provides quantitative decision support for preventive protection of printed matter.

Description

Technical Field

[0001] This invention relates to the field of image processing and pattern recognition technology, specifically to a method for monitoring printed images based on optical recognition technology. Background Technology

[0002] With increasing awareness of cultural heritage protection and the accelerated digitization of museum collections, the need for long-term, non-destructive, and intelligent condition monitoring of printed materials is becoming increasingly urgent. Traditional visual observation and periodic sampling are insufficient for precise and quantitative preventative protection, necessitating an intelligent monitoring method capable of extracting intrinsic material condition information from optical images and predicting its changing trends.

[0003] Image monitoring technology for printed materials is mainly used in fields such as digital archiving of documents, preservation of ancient books and archives management. Among existing technologies, optical character recognition technology achieves digitization by recognizing text content, but this method only focuses on "what the content is" and cannot perceive the structural state of the paper itself or the chemical changes of the ink. Image defect detection technology locates obvious defects such as stains and tears by comparing the differences between the current image and the standard image, but such methods are post-event detection, are sensitive to lighting conditions, and are difficult to achieve quantitative comparison of data from different periods, and cannot assess the early deterioration trend of materials.

[0004] Another type of existing technology is based on hyperspectral imaging for material analysis, which identifies the composition of substances by analyzing spectral curves. However, the equipment is expensive and the data processing is complex, making it difficult to achieve long-term, large-scale, and automated daily monitoring. Existing technologies generally lack the comprehensive ability to convert optical signals into material health indicators with clear physical meaning and to calculate degradation rates and predict remaining life based on time-series data. They cannot truly serve the needs of preventive protection for refined quantitative assessment of material status and forward-looking decision-making. Summary of the Invention

[0005] To address the aforementioned shortcomings of existing technologies, this invention provides a printed image monitoring method based on optical recognition technology, which can effectively solve the problem in the prior art of being unable to quantitatively assess material state and predict trends.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a method for monitoring printed matter images based on optical recognition technology, comprising:

[0007] S1. Collect the multimodal optical features corresponding to the printed matter to be identified, and generate the corresponding standardized multimodal dataset.

[0008] The standardized multimodal dataset includes visible light images, near-infrared images, ultraviolet fluorescence images, and surface normal maps.

[0009] S2. Divide the visible light image into regions, calculate the comprehensive health index of paper structure and the quantitative index of ink photochemical properties of the corresponding region of the printed matter to be identified, and generate the spatial distribution map of the printed matter to be identified.

[0010] S3. Calculate the changes in paper structure density index and ink chemical stability index corresponding to the printed matter to be identified, calculate the feature degradation rate corresponding to the printed matter to be identified, and assess whether risk intervention is required.

[0011] S4. Analyze the remaining lifespan of the printed material to be identified.

[0012] S5. Generate a comprehensive health diagnosis view corresponding to the printed material to be identified.

[0013] Preferably, the process of dividing the visible light image into corresponding regions is as follows:

[0014] The visible light image corresponding to the printed matter to be identified is obtained, and optical recognition is performed on the visible light image to segment the text area, blank area and edge area of ​​the visible light image corresponding to the printed matter to be identified.

[0015] Preferably, the specific process for calculating the paper structure density index of the corresponding area of ​​the printed matter to be identified is as follows:

[0016] Based on the text area, blank area, or edge area of ​​the visible light image corresponding to the printed matter to be identified, for each pixel in the image, when the pixel belongs to the blank area or the edge area, a square local neighborhood with a set side length is defined with the area as the center.

[0017] By combining the total number of pixels in the local neighborhood of the square and the coordinates of each pixel, as well as the grayscale value of each pixel in the local neighborhood of the square in the near-infrared image, the near-infrared local mean value of the printed matter to be identified is calculated.

[0018] Obtain the normal vectors of each pixel within the local neighborhood of the square corresponding to the printed material to be identified. Combine this with the total number of pixels in the local neighborhood of the square to calculate the average vector of all normal vectors within the neighborhood corresponding to the printed material to be identified.

[0019] Obtain the angle between the normal of each pixel and the average normal, and calculate the root mean square of the corresponding angle by combining the total number of pixels in the local neighborhood of the square.

[0020] By combining the near-infrared local mean value in the local neighborhood of the square corresponding to the printed matter to be identified, the root mean square of the angle between the normal of each pixel and the average normal, and the set minimum constant, the paper structure density index of the region corresponding to the printed matter to be identified is calculated.

[0021] Preferably, the ink chemical stability index of the corresponding area of ​​the printed matter to be identified is calculated, and the specific process is as follows:

[0022] Based on each pixel in the image corresponding to the printed matter to be identified, when the pixel belongs to the text area, the gray values ​​of the red channel, green channel, and ultraviolet fluorescence image corresponding to the pixel are obtained, and the ink chemical stability index of the area corresponding to the printed matter to be identified is calculated.

[0023] Preferably, the specific process for generating the spatial distribution map corresponding to the printed matter to be identified is as follows:

[0024] Based on the original visible light image corresponding to the printed matter to be identified, a first blank matrix and a second blank matrix of the same size are created.

[0025] Based on the paper structure density index corresponding to each pixel, each pixel in the image corresponding to the printed matter to be identified is traversed. If the pixel belongs to a blank area or an edge area, the calculated paper structure density index of the pixel is filled into the corresponding position in the first blank matrix. If the pixel belongs to a text area, the corresponding position in the first blank matrix is ​​marked as invalid.

[0026] Based on the ink chemical stability index corresponding to each pixel, each pixel in the image corresponding to the printed matter to be identified is traversed. If the pixel belongs to the text area, the ink chemical stability index corresponding to the pixel is filled into the corresponding position of the second blank matrix. If the pixel belongs to the blank area or the edge area, it is marked as invalid in the corresponding position of the second blank matrix. The filled numerical spectrum is then converted into a pseudo-color image to generate the spectrum corresponding to the printed matter to be identified.

[0027] Preferably, the specific process for calculating the change in paper structure density index and the change in ink chemical stability index corresponding to the printed matter to be identified is as follows:

[0028] Obtain the paper structure density index and ink chemical stability index of each pixel in the image corresponding to the printed matter to be identified at each time stamp, and calculate the change in the paper structure density index and the change in the ink chemical stability index of each pixel corresponding to the printed matter to be identified at the current time stamp.

[0029] Based on the changes in the paper structure density index and the ink chemical stability index of each pixel at the current timestamp, a change field matrix of the same size as the original image of the printed matter to be identified is generated.

[0030] Preferably, the specific process for calculating the feature degradation rate corresponding to the printed matter to be identified is as follows:

[0031] Based on the current timestamp and historical adjacent timestamps corresponding to the changes in the paper structure density index and the ink chemical stability index of the printed matter to be identified, the time interval corresponding to the changes is calculated.

[0032] For each of the blank areas, text areas, and edge areas, perform the following steps:

[0033] Step 1: Filter all pixels belonging to a certain region whose corresponding paper structure density index change is negative, construct a set of paper structure density index changes, select the representative value of the paper structure density index change for that region, and calculate the degradation rate of the paper structure density feature for that region by combining the time interval corresponding to the change.

[0034] Step 2: Filter all pixels belonging to the region whose corresponding ink chemical stability index change is negative, construct a set of ink chemical stability index change values, select the representative value of the ink chemical stability index change value for the region, and calculate the ink chemical stability feature degradation rate for the region based on the time interval corresponding to the change value.

[0035] Preferably, the process for assessing whether risk intervention is required is as follows:

[0036] Based on the paper structure density index and ink chemical stability index of each pixel of the printed matter to be identified, and combined with the set first and second safety thresholds, the paper structure risk and ink condition risk of each pixel area are assessed.

[0037] Based on the region to which each pixel belongs, the degradation rate of the paper structure density feature and the degradation rate of the ink chemical stability feature corresponding to that region, and combined with the set degradation rate threshold coefficient and the time interval of change, the maximum negative change threshold of the paper structure and the maximum negative change threshold of the ink chemical stability of the region to which each pixel belongs within the corresponding time interval are calculated respectively.

[0038] Based on the changes in the paper structure density index and the ink chemical stability index corresponding to the printed matter to be identified, and combined with the maximum negative change threshold of the region to which each pixel belongs within the corresponding time interval, the deterioration trend of the region to which each pixel belongs is evaluated to determine whether it meets the requirements.

[0039] The overall risk level of each area of ​​the printed matter to be identified is divided into four levels, with the risk increasing from low to high. No risk intervention is performed only when the overall risk level of the area to which each pixel belongs is the first level.

[0040] Preferably, the process of analyzing the remaining lifetime of the printed matter to be identified is as follows:

[0041] By combining the material state index values ​​and aging rate constants of the printed materials to be identified in each region corresponding to adjacent historical timestamps, the material state index values ​​and time intervals of the printed materials to be identified in each region corresponding to the current timestamp, the material performance degradation of the printed materials to be identified in each region corresponding to the current timestamp is calculated.

[0042] Based on the feature degradation rate of each region of the printed matter to be identified, the feature degradation rate includes the paper structure density feature degradation rate or the ink chemical stability feature degradation rate. For blank areas and edge areas, the paper structure density feature degradation rate is used, and for text areas, the ink chemical stability feature degradation rate is used. Combining the feature degradation rate of each region with the average value of the current state value of all pixels, the aging rate constant corresponding to each region is calculated.

[0043] By combining the current state value of each pixel in each region, the first security threshold and the second security threshold, and the aging rate constant corresponding to each region, the remaining lifespan of each region of the printed matter to be identified is calculated.

[0044] The technical solution provided by this invention has the following advantages compared with the known prior art:

[0045] 1. In the multimodal data acquisition process, this invention simultaneously acquires visible light, near-infrared, ultraviolet fluorescence, and grazing light images, and synthesizes them into a surface normal map, constructing a spatially strictly aligned standardized multimodal dataset. Simultaneously, through environmental temperature and humidity balance control, closed-loop feedback adjustment of light source intensity, and high-precision geometric calibration, the influence of environmental fluctuations and differences in acquisition conditions on the image data is eliminated. This facilitates providing a comparable and repeatable data foundation for all subsequent quantitative analyses, ensuring accurate pixel-level temporal comparisons of images acquired at different times.

[0046] 2. In the material health index calculation process of this invention, the paper and ink indicators are calculated separately in a clean area by dividing the area into regions. The paper structure density index and the ink chemical stability index are designed. The paper structure density index combines the sensitivity of near-infrared to fiber density with the sensitivity of normal map to surface roughness, and completely eliminates the influence of light by using the standard deviation of normal map. The ink chemical stability index combines the difference between red and green channels with the intensity of ultraviolet fluorescence to achieve a multi-dimensional assessment of the chemical state of ink. This is conducive to outputting a state field that is decoupled from the light conditions, has clear physical meaning, and can be directly used for time-series comparison, so that the monitored object is truly transformed from an "image" into a "material state field".

[0047] 3. In the process of time-series change analysis and risk diagnosis, this invention generates a change field by calculating pixel-level differences to locate "where changes have occurred and by how much". Based on this, for each region, all pixels with negative changes are selected, and the 15th percentile is used as the representative value of the change to calculate the feature degradation rate, focusing on the typical deterioration rate of the "poorest 15% region" within the region. Then, a "state-trend" two-dimensional risk judgment model is constructed to divide each region into four risk levels, which helps to eliminate noise interference while accurately capturing early deterioration areas, and outputs differentiated intervention suggestions for different risk areas.

[0048] 4. In the remaining lifetime prediction process, this embodiment of the invention converts the feature degradation rate into an aging rate constant based on the first-order reaction kinetic equation. Combined with the current state value and safety threshold, it calculates the remaining safe years of each pixel and the average remaining lifetime of each region. This helps to establish a scientific correlation between the average rate of change of discrete time intervals and the instantaneous rate constant in the continuous model, and transforms monitoring data into forward-looking early warning information for the future.

[0049] 5. In the process of comprehensive diagnosis and decision output, the embodiments of the present invention use the original visible light image as the background and overlay the current status map, change field, risk level distribution and remaining life prediction results to generate a multi-layer fused comprehensive health diagnosis view. This is conducive to cultural relics management personnel to have a clear understanding of the overall health distribution, recent change characteristics, risk area location and future life prediction of printed materials, and to transform complex quantitative analysis results into decision support information that is easy to understand. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0051] Figure 1 This is a schematic diagram of the implementation steps of the present invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0053] The present invention will be further described below with reference to embodiments.

[0054] Please see Figure 1 As shown, the printed matter image monitoring method based on optical recognition technology includes at least the following:

[0055] S1. Collect the multimodal optical features corresponding to the printed matter to be identified, and generate the corresponding standardized multimodal dataset.

[0056] Multimodal optical features include visible light images, near-infrared images, ultraviolet fluorescence images, and grazing light images.

[0057] In one specific embodiment, the process of collecting the multimodal optical features corresponding to the printed matter to be identified is as follows: the printed matter to be identified is placed in the sample area of ​​the acquisition platform, the sample area has a neutral gray background, the ambient temperature and humidity of the acquisition platform are stabilized within the set temperature and humidity range, and the printed matter to be identified is left to stand for a set time until the printed matter to be identified and the environment of the acquisition platform reach thermal and humidity equilibrium.

[0058] Each independent light source module, including a white LED array, a near-infrared LED array, and an ultraviolet LED array, has its luminous intensity adjusted to a preset nominal value by a built-in sensor during each image acquisition.

[0059] The sample stage carrying the printed material to be identified is subjected to a preset translational motion. The standard calibration plate fixed on the sample stage is photographed from multiple angles by a camera. The camera parameters are calibrated by computer vision algorithms (such as Zhang Zhengyou calibration method), thereby establishing a mapping relationship between the image pixel coordinates and the physical coordinates of the sample stage.

[0060] After calibration, fix the relative position of the printed material to be identified and the camera, and perform data acquisition sequentially:

[0061] Turn on the white LED array to capture visible light color images of the printed material to be identified.

[0062] Turn off the white LED array, turn on the near-infrared LED array, acquire the near-infrared reflected image of the printed material to be identified, and turn off all light source modules.

[0063] Turn on the ultraviolet LED array and acquire ultraviolet-induced fluorescence images of the printed material to be identified.

[0064] With all light source modules turned off, a white LED light source is rotated at a set angle around the center of the printed matter to be identified. The white LED light source is then stopped at equal intervals along the circumference and images of the printed matter to be identified are captured at the corresponding positions, thereby obtaining a grazing light image sequence corresponding to the printed matter to be identified.

[0065] The visible light image, near-infrared image, ultraviolet fluorescence image and grazing light image corresponding to the printed matter to be identified are preprocessed, and the grazing light image sequence is synthesized into a normal map representing the surface micromorphology through a photometric stereo vision algorithm. Then, a standardized multimodal dataset containing visible light image, near-infrared image, ultraviolet fluorescence image and surface normal map is output.

[0066] It should be noted that the preset nominal value is determined through preliminary system calibration experiments based on the spectral characteristics of the light source, the optical sensitivity of the printed material to be identified, and the dynamic response range of the imaging sensor. In the actual acquisition process, the current luminous intensity is monitored in real time by the built-in light sensor before each acquisition, and the drive current is automatically adjusted by the closed-loop feedback control system to keep the deviation between the actual output intensity and the nominal value within the preset tolerance range (such as within ±1%), thereby eliminating the differences in acquisition conditions caused by light source aging or environmental fluctuations.

[0067] It should be noted that the standard calibration plate refers to a flat plate with high-precision feature patterns (such as checkerboard corner arrays or circular arrays) printed on its surface. The geometric dimensions and spatial positions of the patterns are known in advance and are used to provide a spatial reference for camera calibration.

[0068] It should be noted that the computer vision algorithm mentioned in this scheme refers to the Zhang Zhengyou calibration method, which is used to calculate the camera's internal parameters (such as focal length, principal point coordinates, and distortion coefficients) and external parameters (such as rotation matrix and translation vector) by analyzing multiple images of the calibration board under different poses.

[0069] It should be noted that the camera parameter calibration refers to the process of determining the internal geometric and optical characteristics (intrinsic parameters) of the camera imaging model and the position and attitude of the camera in the world coordinate system (extrinsic parameters) through mathematical methods. For example, by shooting a checkerboard calibration board at different angles, the algorithm calculates the corresponding proportional relationship between each pixel in the image and the actual physical size, while eliminating image distortion caused by lens distortion.

[0070] It should be noted that the mapping relationship is established through the camera intrinsic and extrinsic parameters obtained from calibration, specifically represented by a coordinate transformation matrix. When the pixel coordinates of a point in the image are known (e.g., row 500, column 300), the actual position coordinates of that point in the physical space of the sample stage (e.g., 10 mm in the horizontal direction and 8 mm in the vertical direction) can be calculated in reverse through this matrix.

[0071] It should be noted that the synthesis of grazing light image sequences into surface normal maps is based on the principle of photometric stereo vision. That is, by analyzing the brightness changes of the same pixel in multiple images of the same scene under illumination from different directions, and using known light source direction information and surface reflection models, the normal direction vector of the surface at that point is solved in reverse.

[0072] It should be noted that the standardized multimodal dataset refers to a spatially aligned dataset that includes visible light images, near-infrared images, ultraviolet fluorescence images, and surface micromorphology information (such as normal maps) reconstructed from grazing light image sequences after spatial registration, serving as a standardized input source for subsequent quantitative analysis steps.

[0073] In this embodiment of the invention, during the multimodal data acquisition process, visible light, near-infrared, ultraviolet fluorescence, and grazing light images are simultaneously acquired and synthesized into a surface normal map, constructing a spatially strictly aligned standardized multimodal dataset. Simultaneously, through environmental temperature and humidity balance control, closed-loop feedback adjustment of light source intensity, and high-precision geometric calibration, the influence of environmental fluctuations and differences in acquisition conditions on the image data is eliminated. This facilitates providing a comparable and repeatable data foundation for all subsequent quantitative analyses, ensuring accurate pixel-level temporal comparisons of images acquired at different times.

[0074] S2. Based on the standardized multimodal dataset, the visible light image is divided into regions, and the paper structure comprehensive health index and ink photochemical property quantification index of the corresponding region of the printed matter to be identified are calculated to generate the spatial distribution map of the printed matter to be identified.

[0075] In one specific embodiment, the process of dividing the visible light image into regions is as follows: the visible light image corresponding to the printed matter to be identified is obtained from the standardized multimodal dataset, the visible light image is identified through a semantic segmentation network, and then the text region, blank region and edge region of the visible light image corresponding to the printed matter to be identified are segmented.

[0076] It should be noted that the semantic segmentation network is a pixel-level classification model based on deep learning, which is a mature existing technology in the field of computer vision. In this solution, the U-Net architecture or its variants can be used. The segmentation process is as follows: the visible light image is input into the trained network model, and the network extracts image features layer by layer through the encoder-decoder structure. Finally, it outputs a category probability for each pixel, and classifies the pixel into a text region, blank region or edge region according to the maximum probability.

[0077] In one specific embodiment, the calculation of the paper structure density index and ink chemical stability index of the corresponding area of ​​the printed matter to be identified is carried out as follows: Based on the text area, blank area, or edge area of ​​the visible light image corresponding to the printed matter to be identified, for each pixel in the image, when the pixel belongs to the blank area or edge area, a square local neighborhood with a set side length is defined with the area as the center. , and The horizontal and vertical axes, respectively, represent the pixels of the image corresponding to the printed matter to be identified, calculated using the mean formula: The near-infrared local mean value corresponding to the printed matter to be identified is calculated. ,in and Represented as local neighborhoods of squares The x and y axes of the pixel coordinates within the cell. Represented as the total number of pixels in the local neighborhood of the square. Represented as pixels in the corresponding square local neighborhood in the near-infrared image of the printed matter to be identified. The grayscale value.

[0078] Obtain the normal vectors of each pixel within the local neighborhood of the square corresponding to the printed material to be identified. Through the calculation formula: The average vector of all normal vectors in the neighborhood corresponding to the printed matter to be identified is calculated. .

[0079] Obtain the angle between the normal of each pixel and the average normal. Through the calculation formula: The root mean square of the corresponding included angle is calculated. .

[0080] Combining the local mean near-infrared value of the printed material to be identified and the root mean square of the angle between the normal of each pixel and the average normal, the following formula is used for calculation: The paper structure density index of the corresponding area of ​​the printed matter to be identified was calculated. ,in For a set minimum constant, This is represented as the root mean square reference value of the angle between the surface normals of the standard flat paper.

[0081] Based on each pixel in the image corresponding to the printed matter to be identified, when the pixel belongs to the text area, the grayscale value of the red channel corresponding to that pixel is obtained. Green channel grayscale value and ultraviolet fluorescence image grayscale value Through the calculation formula: The ink chemical stability index of the corresponding area of ​​the printed matter to be identified was calculated. .

[0082] It should be noted that defining a square local neighborhood with a set side length centered on this region refers to the area centered on the current pixel. Centered on, let the side length be... Expanding to both sides along the horizontal and vertical axes on the image plane. , forming a collection A square window of pixels is defined, and all pixels within this window constitute a local neighborhood for subsequent statistical calculations. .

[0083] It should be noted that the acquisition of each normal vector is based on the standardized multimodal dataset output by S1. From this dataset, a surface normal map pre-synthesized from the grazing light image sequence using a photometric stereo vision algorithm is extracted. This normal map is a three-channel image strictly registered with the visible light image, with each pixel storing a three-dimensional unit normal vector. During calculation, for the current pixel... Based on its coordinates, the normal vector values ​​corresponding to all pixels in its local neighborhood are directly read from the normal map and used as input for subsequent average vector calculation.

[0084] It should be noted that the angle between the normal vector of each pixel and the average normal vector is obtained as follows: based on the normal vectors within the local neighborhood of the square corresponding to the printed material to be identified and the average vector of all normal vectors within the neighborhood, the angle is calculated using the formula: The angle between the normal of each pixel and the average normal is calculated. .

[0085] It should be noted that, The setup method is as follows: Select a paper sample of the same material as the printed matter to be identified, which is known to be well-preserved and has a smooth surface. Measure its surface normal map under standardized acquisition conditions, calculate the root mean square (RMS) of the angle between the normals corresponding to all pixels and the average normal, and take the statistical average (such as the mean or median) of these RMS values ​​as... .

[0086] It should also be noted that the acquisition of the gray values ​​of the red channel, green channel, and ultraviolet fluorescence image corresponding to the pixel is based on a standardized multimodal dataset from which visible light color image and ultraviolet fluorescence image are extracted. The two have been pre-registered to the same spatial coordinate system. For the current pixel, based on the region division result, it is determined whether the current pixel belongs to the text region. Then, the gray value of the pixel is directly read from the red channel and green channel of the visible light image, and the gray value is read from the corresponding coordinate position of the ultraviolet fluorescence image.

[0087] In one specific embodiment, the process of generating the spatial distribution map corresponding to the printed matter to be identified is as follows: based on the original visible light image corresponding to the printed matter to be identified, a first blank matrix and a second blank matrix of the same size are created.

[0088] Based on the paper structure density index corresponding to each pixel, each pixel in the image corresponding to the printed matter to be identified is traversed. If the pixel belongs to a blank area or an edge area, the calculated paper structure density index of the pixel is filled into the corresponding position in the first blank matrix. If the pixel belongs to a text area, the corresponding position in the first blank matrix is ​​marked as invalid.

[0089] Based on the ink chemical stability index corresponding to each pixel, each pixel in the image corresponding to the printed matter to be identified is traversed. If the pixel belongs to the text area, the ink chemical stability index corresponding to the pixel is filled into the corresponding position of the second blank matrix. If the pixel belongs to the blank area or the edge area, it is marked as invalid in the corresponding position of the second blank matrix. The filled numerical map is converted into a pseudo-color image to generate the map corresponding to the printed matter to be identified. The generated image and the corresponding metadata are encapsulated and stored.

[0090] It should be noted that the first blank matrix and the second blank matrix are two two-dimensional arrays filled with all zeros or null values, created in computer memory based on the size (height and width in pixels) of the original visible light image corresponding to the printed matter to be identified. Each array element corresponds one-to-one with the pixels of the original image.

[0091] It should be noted that converting the filled numerical spectrum into a pseudo-color image means converting the value of each pixel in the two generated numerical spectra (paper structure density index spectrum and ink chemical stability index spectrum) into the corresponding RGB color value through a preset color mapping function (such as a linear or piecewise function). Usually, a blue-green-red gradient color system is used so that low value areas appear blue, medium value areas appear green, and high value areas appear red.

[0092] It should also be noted that the metadata includes the generation timestamp, the numerical range of the paper structure density index and the ink chemical stability index, the color mapping relationship and spatial information. The spatial information refers to the correspondence between the map and the physical space of the printed matter to be identified, including the image resolution (the actual physical size corresponding to each pixel, such as millimeters / pixel), the width and height of the map in pixels, and the coordinate system information shared by the map and the original visible light image.

[0093] In the material health index calculation process of this invention, the paper and ink indicators are calculated separately in a clean area by dividing the area into regions. The paper structure density index and the ink chemical stability index are designed. The paper structure density index combines the sensitivity of near-infrared to fiber density with the sensitivity of normal map to surface roughness, and completely eliminates the influence of light by using the standard deviation of normal map. The ink chemical stability index combines the difference between red and green channels with the intensity of ultraviolet fluorescence to achieve a multi-dimensional assessment of the chemical state of ink. This is conducive to outputting a state field that is decoupled from the light conditions, has clear physical meaning, and can be directly used for time-series comparison, so that the monitored object is truly transformed from an "image" into a "material state field".

[0094] S3. Based on the spatial distribution map of the printed matter to be identified, calculate the change in paper structure density index and ink chemical stability index of the printed matter to be identified, and calculate the feature degradation rate of the printed matter to be identified, so as to assess whether risk intervention is required.

[0095] In one specific embodiment, the calculation of the change in paper structure density index and the change in ink chemical stability index corresponding to the printed matter to be identified is carried out as follows: the paper structure density index and ink chemical stability index of each pixel in the image corresponding to the printed matter to be identified are obtained from the database at each time stamp; the paper structure density index and ink chemical stability index of adjacent historical time stamps are subtracted from the paper structure density index and ink chemical stability index corresponding to the current time stamp to obtain the change in paper structure density index and the change in ink chemical stability index of each pixel corresponding to the printed matter to be identified at the current time stamp.

[0096] If the change in the paper structure density index of a pixel corresponding to the printed matter to be identified at the current timestamp is positive, it indicates that the paper condition of the area to which the pixel belongs has improved; if it is negative, it indicates that the paper condition has deteriorated; and if it is zero, it indicates that the paper condition is stable.

[0097] If the change in the ink chemical stability index of a pixel corresponding to the printed matter to be identified at the current timestamp is positive, it indicates that the ink stability of the paper in the area to which the pixel belongs has improved; if it is negative, it indicates that the ink chemical stability has decreased.

[0098] Based on the changes in the paper structure density index and the ink chemical stability index of each pixel of the printed matter to be identified at the current timestamp, a change field matrix of the same size as the original image of the printed matter to be identified is generated, including the paper structure density index change field matrix and the ink chemical stability index change field matrix.

[0099] It should be noted that the sign and magnitude of the change in the paper structure density index directly reflect the direction and degree of change in the paper's microstructure. If the change is positive, it indicates that the paper structure in that area tends to be denser and the surface tends to be smoother, for example, due to fiber shrinkage caused by reduced environmental humidity, resulting in increased paper compactness, or due to long-term pressing causing surface wrinkles to smooth out. If the change is negative, it indicates that the paper structure in that area tends to be looser and the surface roughness increases, for example, due to fiber expansion and loosening caused by moisture, fiber breakage caused by acidification, or surface fuzzing caused by mechanical wear. If the change is zero, it indicates that the paper structure in that area remains relatively stable during the two monitoring intervals.

[0100] It should be noted that the sign and magnitude of the change in the ink chemical stability index directly reflect the direction and extent of the change in the chemical state of the ink. A positive change indicates that the chemical stability of the ink in that area has relatively improved. This could be due to the removal of temporary surface contaminants (such as mold or dust) leading to a decrease in ultraviolet fluorescence, or due to environmental improvements slowing down the ink oxidation process, thus resulting in a restorative increase in the ICSI value. Conversely, a negative change indicates that the chemical stability of the ink in that area has decreased. This could be due to continuous ink oxidation leading to the formation of fluorescent substances, the production of fluorescent secretions by microbial metabolism, ink diffusion leading to a decrease in the red-green channel difference, or the aging and degradation of modern repair ink.

[0101] It should be noted that the generated change field matrix refers to filling the paper structure density index change and ink chemical stability index change of each pixel at the current timestamp into two blank matrices of the same size as the original image, according to the coordinate position of the pixel in the original image. Each row and column of the matrix strictly corresponds to each row and column of the original image, thus forming two complete two-dimensional arrays, which are called the paper structure density index change field matrix and the ink chemical stability index change field matrix, respectively.

[0102] It should be noted that, due to the instantaneous interference such as sensor noise, subpixel-level registration error or tiny surface contaminants that may exist during image acquisition, the change in a single pixel may be random. In actual analysis, this scheme does not rely solely on the value of a single pixel for judgment, but combines the local neighborhood statistics (which already includes the spatial smoothing effect) used in S2 to calculate the index, the calculation of the regional feature degradation rate in S3 (taking percentiles rather than single pixels), and the consistency verification of multi-period time series data.

[0103] In one specific embodiment, the process of calculating the feature degradation rate corresponding to the printed matter to be identified is as follows: based on the current timestamp and historical adjacent timestamps corresponding to the change in the paper structure density index and the change in the ink chemical stability index corresponding to the printed matter to be identified, the time interval corresponding to the change is calculated.

[0104] Based on the blank area, text area, and edge area corresponding to the printed material to be identified, perform the following steps for each of the blank area, text area, and edge area:

[0105] Step 1: Select all pixels belonging to the region and whose corresponding paper structure density index change is negative from the paper structure density index change field matrix. Construct a paper structure density index change set from the paper structure density index change values ​​of each pixel that meets the condition. Sort the elements in the set in ascending order of value and take the 15th percentile as the representative value of the paper structure density index change for the region. Divide the representative value of the paper structure density index change for the region by the time interval corresponding to the change to obtain the degradation rate of the paper structure density feature for the region.

[0106] Step 2: Filter out all pixels belonging to the region and whose corresponding ink chemical stability index change is negative from the ink chemical stability index change field matrix. Construct an ink chemical stability index change set from the ink chemical stability index change values ​​of each pixel that meets the condition. Sort the elements in the set in ascending order of value, and take the 15th percentile as the representative value of the ink chemical stability index change for the region. Divide the representative value of the ink chemical stability index change for the region by the time interval corresponding to the change to obtain the ink chemical stability characteristic degradation rate for the region.

[0107] In one specific embodiment, the assessment of whether risk intervention is required involves the following process: Based on the paper structure density index of each pixel corresponding to the printed matter to be identified, a determination is made for each pixel belonging to a blank area or an edge area: ,in This is represented as the set first safety threshold. Represented as pixels In the blank area, Represented as pixels Located in the peripheral area, Represented as pixels Is there a binary identifier for paper structure risk? , representing pixels If the paper condition is below the first safety threshold, the area containing this pixel poses a paper structure risk. Represents pixels, There is no paper structure risk in the area. This indicates a logical OR.

[0108] Based on the ink chemical stability index of each pixel in the printed material to be identified, the determination is made for each pixel belonging to the text area: ,in Represented as pixels Is there a binary indicator indicating whether there is a risk related to ink condition? and Represented as pixels The area may or may not have ink condition risks. This is represented as the set second security threshold. Represented as pixels This is a text area. This indicates a logical AND operation.

[0109] Based on pixels The degradation rate of the paper structure density characteristics of the region and the corresponding region and the degradation rate of ink chemical stability characteristics Through the calculation formula: and The pixels were calculated separately. The region belongs to the time interval The threshold for the maximum negative change in the paper structure corresponding to the internal structure. Threshold for maximum negative change in ink chemical stability ,in This is represented as the set degradation rate threshold coefficient. This represents the time interval of the change calculated in S3.

[0110] Based on the change in the density index of the paper structure corresponding to the printed matter to be identified Changes in ink chemical stability index Combined with pixels The region belongs to the time interval The corresponding maximum negative change threshold is determined by the risk assessment formula: and , evaluate pixels Does the deteriorating trend in the region meet the requirements? or When, it indicates the pixel The deterioration trend in the region does not meet the requirements; conversely, it does meet the requirements if the trend is positive.

[0111] The overall risk level of the printed materials to be identified for each region is divided into four levels, with the risk increasing sequentially from low to high. =0 and When it is 0, the pixel The overall risk level of the area is Level 1.

[0112] when =0 and When it is 1, the pixel The overall risk level of the area is Level 2.

[0113] when =1 and When it is 0, the pixel The overall risk level of the area is Level 3.

[0114] when =1 and When it is 1, the pixel The overall risk level of the area is Level 4.

[0115] If and only if pixel No risk intervention will be carried out when the overall risk level of the region is Level 1.

[0116] It should be noted that, include and , 0 means as and All are 0. include and , 0 indicates and All are 0. For 1 and The meaning of 1 is the same as above.

[0117] It should be noted that the process of setting the first safety threshold is based on the correlation test calibration of paper mechanical strength and paper structure density index. Specifically, aging samples of the same material as the printed matter to be identified are selected, and their tensile strength (or folding endurance) and paper structure density index are measured simultaneously in a laboratory environment. A fitting curve between the two is established, and the paper structure density index value when the strength drops to 50% of the initial value or the recognized safety critical value (such as according to the cultural relics protection industry standard) is determined as the first safety threshold.

[0118] It should be noted that the setting process of the second safety threshold is based on the statistical distribution analysis of the ink chemical stability index of known stable inks. Stable ink samples of the same era and type as the printed matter to be monitored and confirmed to have not undergone significant chemical changes are collected. The mean and standard deviation of their ink chemical stability index values ​​are calculated, and the second safety threshold is set as the mean minus 2 times the standard deviation (or adjusted according to the confidence level requirements).

[0119] It should be noted that the degradation rate threshold coefficient is set based on a combination of statistical principles and empirical calibration. By collecting a large amount of historical monitoring data of similar printed materials, the distribution of the characteristic degradation rate of each region is calculated. The upper quartile (i.e., the 75th quartile) or the mean plus one standard deviation of this distribution is taken as the upper limit of "normal degradation". Then, this upper limit value is divided by the characteristic degradation rate of each region to obtain a preliminary coefficient reference value. On this basis, it is fine-tuned through experimental verification or expert experience (usually between 1.5 and 2.0) to obtain the final degradation rate threshold coefficient.

[0120] It should be noted that for the first risk level, i.e., areas in good and stable condition, maintaining the existing monitoring frequency (e.g., once a year) is sufficient, and no special intervention is required; for the second risk level, i.e., areas in acceptable condition but deteriorating rapidly, it is recommended to shorten the monitoring cycle (e.g., to once every six months) and closely monitor its changing trends; for the third risk level, i.e., areas in poor condition but relatively stable recently, it is recommended to include them in the next maintenance cycle plan and arrange local environmental optimization or preventive treatment; for the fourth risk level, i.e., areas in poor condition and still deteriorating rapidly, the specific coordinates, area and current indicator values ​​of the area must be output immediately, and targeted protective measures, such as local reinforcement, deacidification treatment or microenvironment control, must be initiated.

[0121] In the process of time-series change analysis and risk diagnosis, this invention generates a change field by calculating pixel-level differences to locate "where changes have occurred and by how much". Based on this, for each region, all pixels with negative changes are selected, and the 15th percentile is used as the representative value of the change to calculate the feature degradation rate, focusing on the typical deterioration rate of the "poorest 15% of regions" within the region. Then, a two-dimensional risk judgment model of "state-trend" is constructed to divide each region into four risk levels, which helps to eliminate noise interference while accurately capturing early deterioration areas, and outputs differentiated intervention suggestions for different risk areas.

[0122] S4. Analyze the remaining lifespan of the printed material to be identified.

[0123] In one specific embodiment, the process of analyzing the remaining lifetime of the printed matter to be identified is as follows:

[0124] By combining the material state index values ​​and aging rate constants of the printed materials to be identified in each region corresponding to adjacent historical timestamps, the material state index values ​​of the printed materials to be identified in each region corresponding to the current timestamp, and the time interval, the first-order reaction kinetic equation is used to calculate the material performance degradation in each region corresponding to the printed materials to be identified at the current timestamp. .

[0125] Based on the feature degradation rate of each region of the printed matter to be identified, including the degradation rate of paper structure density features or the degradation rate of ink chemical stability features, the paper structure density feature degradation rate is used for blank and edge regions, while the ink chemical stability feature degradation rate is used for text regions. This is expressed by the equation: and The derivation leads to: ,in Represented as The average state value within, It is expressed as the aging rate constant.

[0126] Based on the derivation results, the calculation formula is used: , obtain the area Corresponding aging rate constant ,in Represented as a region The corresponding feature degradation rate, Represented as a region The average of the current state values ​​of all pixels within the range, including the paper structure density index or the ink chemical stability index.

[0127] Based on the first-order reaction kinetic equation, the calculation formula is as follows: The corresponding area of ​​the printed matter to be identified is calculated. Medium pixel Remaining lifespan ,in Represented as a region Medium pixel The current state value, Represented as a region The safety threshold in the text refers to either the first safety threshold or the second safety threshold. If the area... If the region is a text area, then the second security threshold is applied; if the region... If the area is blank or at the edge, the first safety threshold is used.

[0128] according to Computational area The average remaining lifetime of all pixels in the image is used to obtain the remaining lifetime of each region of the printed material to be identified.

[0129] It should be noted that the first-order reaction kinetic equation is a classic model describing the exponential decay of material properties over time. Its core assumption is that the aging rate of the material is proportional to the current state value. This solution applies this model to the prediction of the lifespan of printed materials. By fitting historical monitoring data, the aging rate constant is obtained, and then the time required for the current state value to decay to the safety threshold is extrapolated.

[0130] It should be noted that the approximation symbol “≈” is used in the above derivation because the feature degradation rate is an average change calculated based on a finite time interval and is an instantaneous parameter in a continuous model. The two are approximately equal when the aging rate constant is sufficiently small within a finite time interval. In practical applications, when the monitoring period is relatively short compared to the aging process, this approximation has acceptable accuracy.

[0131] In the remaining lifetime prediction process, this invention uses a first-order reaction kinetic equation to convert the feature degradation rate into an aging rate constant. Combined with the current state value and safety threshold, it calculates the remaining safe years for each pixel and the average remaining lifetime for each region. This helps to establish a scientific correlation between the average rate of change of discrete time intervals and the instantaneous rate constant in the continuous model, and transforms monitoring data into forward-looking early warning information for the future.

[0132] S5. Based on the spatial distribution map, change field, and comprehensive risk level classification results of the printed matter to be identified at the current timestamp, as well as the remaining lifespan of each region, generate a comprehensive health diagnosis view corresponding to the printed matter to be identified.

[0133] In a specific embodiment, the process of generating a comprehensive health diagnosis view corresponding to the printed matter to be identified is as follows: Based on the current state map (spatial distribution map of paper structure comprehensive health index / ink photochemical property quantification index) generated in S2, the change field matrix (change field matrix of paper structure comprehensive health index / change field matrix of ink photochemical property quantification index) generated in S3, the comprehensive risk level classification results, and the remaining lifespan of each region corresponding to the printed matter to be identified in S4, this map uses the original visible light image as a background and overlays the above-mentioned information to intuitively present the overall health distribution, recent change hotspots, risk level areas, and future lifespan expectations.

[0134] In the process of comprehensive diagnosis and decision output, this invention uses the original visible light image as a background to overlay and display the current status map, change field, risk level distribution, and remaining life prediction results, generating a multi-layer fused comprehensive health diagnosis view. This helps cultural relic managers to clearly understand the overall health distribution, recent change characteristics, risk area location, and future life prediction of printed materials, transforming complex quantitative analysis results into easily understandable decision support information.

[0135] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring printed matter images based on optical recognition technology, characterized in that, include: S1. Collect the multimodal optical features corresponding to the printed matter to be identified, and generate the corresponding standardized multimodal dataset; The standardized multimodal dataset includes visible light images, near-infrared images, ultraviolet fluorescence images, and surface normal maps; S2. Divide the visible light image into regions, calculate the comprehensive health index of paper structure and the quantitative index of ink photochemical properties of the corresponding region of the printed matter to be identified, and generate the spatial distribution map of the printed matter to be identified. S3. Calculate the change in paper structure density index and ink chemical stability index of the printed matter to be identified, calculate the feature degradation rate of the printed matter to be identified, and assess whether risk intervention is required. S4. Analyze the remaining lifespan of the printed material to be identified; S5. Generate a comprehensive health diagnosis view corresponding to the printed material to be identified.

2. The printed matter image monitoring method based on optical recognition technology according to claim 1, characterized in that, The specific process for dividing the visible light image into corresponding regions is as follows: The visible light image corresponding to the printed matter to be identified is obtained, and optical recognition is performed on the visible light image to segment the text area, blank area and edge area of ​​the visible light image corresponding to the printed matter to be identified.

3. The printed matter image monitoring method based on optical recognition technology according to claim 2, characterized in that, The specific process for calculating the paper structure density index of the corresponding area of ​​the printed matter to be identified is as follows: Based on the text area, blank area, or edge area of ​​the visible light image corresponding to the printed matter to be identified, for each pixel in the image, when the pixel belongs to the blank area or the edge area, a square local neighborhood with a set side length is defined with the area as the center. By combining the total number of pixels in the local neighborhood of the square and the coordinates of each pixel, as well as the grayscale value of each pixel in the local neighborhood of the square in the near-infrared image, the near-infrared local mean value of the printed matter to be identified is calculated. Obtain the normal vectors of each pixel within the local neighborhood of the square corresponding to the printed material to be identified. Combine this with the total number of pixels in the local neighborhood of the square to calculate the average vector of all normal vectors within the neighborhood corresponding to the printed material to be identified. Obtain the angle between the normal of each pixel and the average normal, and calculate the root mean square of the corresponding angle by combining the total number of pixels in the local neighborhood of the square. By combining the near-infrared local mean value in the local neighborhood of the square corresponding to the printed matter to be identified, the root mean square of the angle between the normal of each pixel and the average normal, and the set minimum constant, the paper structure density index of the region corresponding to the printed matter to be identified is calculated.

4. The printed matter image monitoring method based on optical recognition technology according to claim 3, characterized in that, The specific process for calculating the ink chemical stability index of the corresponding area of ​​the printed material to be identified is as follows: Based on each pixel in the image corresponding to the printed matter to be identified, when the pixel belongs to the text area, the gray values ​​of the red channel, green channel, and ultraviolet fluorescence image corresponding to the pixel are obtained, and the ink chemical stability index of the area corresponding to the printed matter to be identified is calculated.

5. The printed matter image monitoring method based on optical recognition technology according to claim 4, characterized in that, The specific process for generating the spatial distribution map corresponding to the printed matter to be identified is as follows: Based on the original visible light image corresponding to the printed matter to be identified, a first blank matrix and a second blank matrix of the same size are created. Based on the paper structure density index corresponding to each pixel, each pixel in the image corresponding to the printed matter to be identified is traversed. If the pixel belongs to a blank area or an edge area, the calculated paper structure density index of the pixel is filled into the corresponding position in the first blank matrix. If the pixel belongs to a text area, the corresponding position in the first blank matrix is ​​marked as invalid. Based on the ink chemical stability index corresponding to each pixel, each pixel in the image corresponding to the printed matter to be identified is traversed. If the pixel belongs to the text area, the ink chemical stability index corresponding to the pixel is filled into the corresponding position of the second blank matrix. If the pixel belongs to the blank area or the edge area, it is marked as invalid in the corresponding position of the second blank matrix. The filled numerical spectrum is then converted into a pseudo-color image to generate the spectrum corresponding to the printed matter to be identified.

6. The printed matter image monitoring method based on optical recognition technology according to claim 5, characterized in that, The specific process for calculating the changes in the paper structure density index and the ink chemical stability index corresponding to the printed matter to be identified is as follows: Obtain the paper structure density index and ink chemical stability index of each pixel in the image corresponding to the printed matter to be identified at each time stamp, and calculate the change in the paper structure density index and the change in the ink chemical stability index of each pixel corresponding to the printed matter to be identified at the current time stamp. Based on the changes in the paper structure density index and the ink chemical stability index of each pixel at the current timestamp, a change field matrix of the same size as the original image of the printed matter to be identified is generated.

7. The printed matter image monitoring method based on optical recognition technology according to claim 6, characterized in that, The specific process for calculating the feature degradation rate corresponding to the printed matter to be identified is as follows: Based on the current timestamp and historical adjacent timestamps corresponding to the changes in the paper structure density index and the ink chemical stability index of the printed matter to be identified, the time interval corresponding to the changes is calculated. For each of the blank areas, text areas, and edge areas, perform the following steps: Step 1: Filter all pixels belonging to a certain region whose corresponding paper structure density index change is negative, construct a set of paper structure density index change values, select the representative value of the paper structure density index change value of the region, and calculate the degradation rate of the paper structure density feature of the region in combination with the time interval corresponding to the change value. Step 2: Filter all pixels belonging to the region whose corresponding ink chemical stability index change is negative, construct a set of ink chemical stability index change values, select the representative value of the ink chemical stability index change value for the region, and calculate the ink chemical stability feature degradation rate for the region based on the time interval corresponding to the change value.

8. The printed matter image monitoring method based on optical recognition technology according to claim 7, characterized in that, The specific process for assessing whether risk intervention is necessary is as follows: Based on the paper structure density index and ink chemical stability index of each pixel of the printed matter to be identified, and combined with the set first and second safety thresholds, the paper structure risk and ink condition risk of each pixel area are assessed. Based on the region to which each pixel belongs, the degradation rate of the paper structure density feature and the degradation rate of the ink chemical stability feature corresponding to that region, and combined with the set degradation rate threshold coefficient and the time interval of change, the maximum negative change threshold of the paper structure and the maximum negative change threshold of the ink chemical stability of the region to which each pixel belongs within the corresponding time interval are calculated respectively. Based on the changes in the paper structure density index and the ink chemical stability index of the printed matter to be identified, and combined with the maximum negative change threshold of the region to which each pixel belongs within the corresponding time interval, the deterioration trend of the region to which each pixel belongs is evaluated to determine whether it meets the requirements. The overall risk level of each area of ​​the printed matter to be identified is divided into four levels, with the risk increasing from low to high. No risk intervention is performed only when the overall risk level of the area to which each pixel belongs is the first level.

9. The printed matter image monitoring method based on optical recognition technology according to claim 8, characterized in that, The specific process for analyzing the remaining lifetime of the printed matter to be identified is as follows: By combining the material state index values ​​and aging rate constants of the printed materials to be identified in each region corresponding to adjacent historical timestamps, the material state index values ​​of the printed materials to be identified in each region corresponding to the current timestamp, and the time interval, the material performance degradation of the printed materials to be identified in each region corresponding to the current timestamp is calculated. Based on the feature degradation rate of each region of the printed matter to be identified, the feature degradation rate includes the paper structure density feature degradation rate or the ink chemical stability feature degradation rate. For blank areas and edge areas, the paper structure density feature degradation rate is used, and for text areas, the ink chemical stability feature degradation rate is used. Combining the feature degradation rate of each region and the average value of the current state value of all pixels, the aging rate constant corresponding to each region is calculated. By combining the current state value of each pixel in each region, the first security threshold and the second security threshold, and the aging rate constant corresponding to each region, the remaining lifespan of each region of the printed matter to be identified is calculated.

10. The printed matter image monitoring method based on optical recognition technology according to claim 9, characterized in that, The specific process for generating the comprehensive health diagnosis view corresponding to the printed material to be identified is as follows: Based on the spatial distribution map, change field, and comprehensive risk level classification results of the printed matter to be identified at the current timestamp, as well as the remaining lifespan of each region, a comprehensive health diagnosis view corresponding to the printed matter to be identified is generated.

Images