Method and system for detecting defects of micropore structure of low-gram-weight high-strength composite corrugated board

Abstract

The invention provides a low-gram-weight high-strength composite corrugated board micropore structure defect detection method and system, and the method comprises the steps: collecting image data of a corrugated board micropore structure, and constructing a fused corrugated board image set; based on the fused corrugated board image set, the micropore defects in the paperboard are recognized, and a paperboard micropore defect binary mask graph is obtained; spatial relation analysis and structure reconstruction are carried out based on the paperboard micropore defect binary mask graph, and a structure graph of a connected defect area is constructed; analyzing the influence of the micropore defect on the mechanical property of the paperboard material based on the structure diagram of the connected defect area and the geometrical characteristic triple, and predicting the residual strength of the paperboard material under the micropore defect; and based on the residual strength of the paperboard material under the micropore defect, performing micropore defect grade evaluation and hierarchical control strategy formulation. According to the method, detection and performance evaluation of the micropore defects of the corrugated board are achieved, a micropore defect hierarchical control mechanism is constructed, and the method is suitable for the quality control requirements of weight reduction and enhancement of green paperboards.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of packaging material quality detection, and in particular to a low-grammage high-strength composite corrugated paperboard micro-hole structure defect detection method and system. BACKGROUND

[0002] Under the general trend of continuously promoting green low-carbon and high-strength performance material application in the packaging industry, low-grammage high-strength composite corrugated paperboard gradually replaces traditional packaging materials due to its lightweight, high cost-effectiveness, and excellent mechanical properties, and is widely used in many fields such as transportation, warehousing, and e-commerce logistics. The industry has higher requirements for the quality and performance of corrugated paperboard, especially under the background of the widespread use of micro-hole structure design to improve the overall mechanical performance of the material. How to accurately detect micro-hole structure defects and effectively evaluate the mechanical properties of the material has become a key technical bottleneck to ensure the reliability and safety of corrugated paperboard materials.

[0003] Chinese patent application with publication number CN119860983B provides a corrugated paperboard strength detection method and system. The method first obtains the stress data of the corrugated paperboard at each time under different pressing speed tests and the grayscale image of the corrugated side, analyzes the deformation of each corrugated in the grayscale image at each time in each pressing speed test process, obtains the structure damage degree of the corrugated paperboard in each pressing speed test, and extracts the stress peak value of the corrugated paperboard in each pressing speed test. Then, the structure damage degree and stress peak value of each pressing speed test are linearly fitted to obtain the true flat pressure strength of the corrugated paperboard.

[0004] However, the current technology still faces many challenges. At present, the packaging industry generally adopts the design idea of introducing micro-hole structure in the paperboard interlayer of composite corrugated paper to improve the overall mechanical properties of the material. By using microstructure morphology regulation, the stress is effectively dispersed and the impact energy is gradually absorbed, thereby enhancing the compression strength, stiffness, and deformation resistance. However, in actual production, due to the influence of multiple factors such as the uniformity of paper raw materials, the precision control of molding process, and environmental humidity, micro-hole structure is prone to defects such as hole collapse, oversized connected hole size, uneven pore distribution, and interlayer peeling. These subtle defects are usually hidden inside the paperboard or appear as slight structural disturbances, which are difficult to identify through traditional visual inspection methods and can easily become a key cause of structural failure during subsequent use. Especially under low-grammage design conditions, such microstructure defects will directly weaken the high-strength carrying capacity of the paperboard, which is contrary to the design goal of "lightweight and high strength", and seriously affects the stability and safety of the product in complex transportation and stacking scenarios.

[0005] In addition, on the automated high-speed pressing production line, the unit time output of paperboard is large, the rhythm is fast, and once the continuity structure defect occurs in the production process, it is extremely likely to form batch quality hidden trouble in a short time. If the defect is not discovered in time and misflows into the downstream packaging and circulation link, it is extremely likely to induce microstructure stress concentration under extreme working conditions such as moisture, high pressure and impact, and then cause paperboard collapse, packaging damage and other failure events, causing goods damage, logistics interruption, and even serious consequences such as personnel injury, which not only violates the sustainable development concept of green packaging, but also poses a major threat to the reputation and economic benefits of enterprises.

[0006] The existing mainstream quality detection means still relies on manual visual inspection and offline destructive testing, which has the obvious disadvantages of low detection efficiency, strong subjectivity and inability to realize real-time monitoring of the whole process. Especially in the face of high-throughput production tasks, microscopic defects are easy to be ignored, and quality hidden troubles are difficult to be checked in time. SUMMARY

[0007] In order to overcome the above-mentioned defects of the prior art, the present application provides a low grammage high-strength composite corrugated paperboard micro-pore structure defect detection method and system. The method first uniformly cuts, cleans and fixes the composite corrugated paperboard sample to obtain a standardized sample of the composite corrugated paperboard; then configures the microscope acquisition parameters, adopts a multi-focus microscopic imaging technology to collect images layer by layer at different focal heights; then adopts a maximum gray scale fusion algorithm to fuse the multi-focus composite corrugated paperboard images to construct a set of fused high-resolution composite corrugated paperboard images.

[0008] Then, based on the set of fused high-resolution paperboard images, gray scale normalization and denoising preprocessing are performed to obtain a second set of composite corrugated paperboard images; subsequently, a trained convolutional neural network model is used to predict the micro-pore defect prediction of the second set of composite corrugated paperboard images to generate a composite corrugated paperboard micro-pore defect probability map; according to the set threshold, each pixel defect probability value in the composite corrugated paperboard micro-pore defect probability map is converted into a binary mask to obtain a composite corrugated paperboard micro-pore defect binary mask map.

[0009] Further, based on the composite corrugated paperboard micro-pore defect binary mask map, a connectivity clustering analysis of the defect area is performed to identify all independent connected defect areas; and the area, perimeter, equivalent circle diameter and geometric center coordinates of each connected defect area are calculated; at the same time, according to the geometric features of the connected defect area and its defect geometric center set, a triangular subdivision algorithm is used to construct a structure diagram of the connected defect area.

[0010] Subsequently, based on the structural diagram and geometric feature triples of the connected defect area, the weighted porosity of the cardboard material under the microporous defect is calculated; the strength attenuation factor of the cardboard material under the microporous defect is derived through a nonlinear model using the weighted porosity; combined with the initial strength of the cardboard material, the residual strength of the cardboard material under the microporous defect is predicted, realizing a quantitative evaluation of the effect of the defect on the mechanical properties of the cardboard material.

[0011] Finally, based on the residual strength of the cardboard material under microporous defects, the microporous defects are graded based on three preset strength grading thresholds; combined with the process experience rule library and machine learning prediction model, process adjustment plans for different defect levels are automatically generated and integrated; the parameters of the process adjustment plan are legally verified and necessary corrections are made, and finally the final set of process parameter adjustment recommendations is output to guide the production site to specifically adjust key process parameters such as spraying amount, pressing pressure, drying temperature, etc., so as to control the impact of defects and maintain the stable quality of composite corrugated cardboard.

[0012] To achieve the above object, the present invention provides the following technical solutions: A method for detecting microporous structural defects in low-weight, high-strength composite corrugated paperboard, comprising: Collect high-quality image data of the microporous structure of low-weight, high-strength composite corrugated cardboard and construct a fused high-resolution composite corrugated cardboard image set ; Based on the fusion of high-resolution composite corrugated cardboard image collection , identify micropore defects in composite corrugated cardboard and obtain the binary mask image of micropore defects in composite corrugated cardboard ; Based on the binary mask image of micropore defects in composite corrugated cardboard , perform spatial relationship analysis and structural reconstruction, and construct a structural diagram of the connected defect area ; Structural graph based on connected defect regions and the geometric feature triples of connected defect regions , analyze the effect of microporous defects on the mechanical properties of cardboard materials, and predict the residual strength of cardboard materials under microporous defects ; Residual strength of paperboard materials based on microporous defects , conduct micropore defect grade assessment and formulate graded control strategies.

[0013] Furthermore, the constructed fused high-resolution composite corrugated cardboard image set The steps include: Preparation of composite corrugated cardboard samples to obtain standardized samples of composite corrugated cardboard ; Standardized samples based on composite corrugated cardboard , configure the relevant parameters of the microscope collection, complete the microscopic imaging collection at different focal height, and obtain a multi-focus composite corrugated paper image ; Based on the multi-focus composite corrugated paper image , a maximum gray scale fusion algorithm is used to construct a fused high-resolution composite corrugated paper image set ; The composite corrugated paper sample includes but is not limited to paperboard batch number information for identifying different production batch samples; paperboard grammage for measuring the mass per unit area of the paperboard; forming process parameters including hot pressing temperature, pressing strength, forming time and other key process control indexes.

[0014] Further, the step of obtaining the composite corrugated paper micro-hole defect binary mask image includes: Based on the fused high-resolution composite corrugated paper image set , a second composite corrugated paper image set is obtained ; Based on the second composite corrugated paper image set , a convolutional neural network model is used to generate a composite corrugated paper micro-hole defect probability map ; Based on the composite corrugated paper micro-hole defect probability map , threshold segmentation and binary defect region extraction are performed to obtain a composite corrugated paper micro-hole defect binary mask image ; The threshold segmentation and binary defect region extraction are based on a preset threshold parameter, which discriminates the defect probability value of each pixel in the micro-hole defect probability map and performs binary processing.

[0015] Further, the step of spatial relationship analysis and structure reconstruction includes: Based on the composite corrugated paper micro-hole defect binary mask image , a defect region spatial connectivity clustering analysis is performed to extract a set of all independent connected defect sub-regions ; Based on the set of all independent connected defect sub-regions , quantitative analysis of geometric features and extraction of defect geometric center coordinates are performed to construct a geometric feature triple of connected defect regions and a set of defect geometric centers ; Based on the geometric features of the connected defect regions and the set of defect geometric centers , a structure diagram of the connected defect regions is constructed .

[0016] Furthermore, the spatial connectivity cluster analysis of the defect area is to use the 8-neighborhood connectivity rule to analyze the binary mask image of the micropore defect of the composite corrugated cardboard. All foreground pixels (pixel value is 1) are connected; The 8-neighborhood connectivity judgment specifically includes: if a foreground pixel and any of its adjacent neighboring pixels in the four basic directions of up, down, left, and right, or any of its adjacent pixels in the four diagonal directions of upper left, upper right, lower left, and lower right are also foreground pixels, then the two are considered to be the same connected area.

[0017] Furthermore, the geometric feature triples of the connected defect regions are constructed and defect geometric center set The steps include: Traverse the set of all independent connected defect sub-regions in sequence Each defect sub-region in , for each defect sub-region Calculate geometric features separately; Based on the geometric features, a geometric feature triplet of connected defect areas is constructed and the defect geometric center set .

[0018] Furthermore, the step of calculating the geometric features includes: Count each connected defect sub-region The total number of all foreground pixels in , calculate the area of ​​each connected defect sub-region , and after the image is spatially calibrated, the area of ​​each connected defect sub-region is Convert the actual physical area of ​​each connected defect sub-region to obtain the corresponding ; Use boundary extraction algorithm to extract each connected defect sub-region The boundary pixel set , and based on the accumulation of Euclidean distances between adjacent boundary points, the perimeter of each connected defect sub-region is calculated ; Each defect area is approximated as an ideal circular area of ​​equal area, and the equivalent circle diameter of each connected defect sub-area is calculated accordingly ; Based on each connectivity defect sub-region The foreground pixels in the , respectively count the average value of the horizontal and vertical coordinates of each foreground pixel point, extract the geometric center coordinates of each connected defect sub-region .

[0019] Furthermore, the structure diagram of the connected defect area is constructed The steps include: Based on each connectivity defect sub-region The geometric center coordinates of , is regarded as a node of the structure graph, and each graph node is bound to the corresponding geometric feature triple as a node attribute to form a graph node , build a complete set of graph nodes ; Based on graph node collection Coordinates of the geometric centers , use Delaunay triangulation algorithm to establish the connection relationship. Center point 、 If they are adjacent in the Delaunay grid, an undirected edge is established. , construct a complete graph edge set ; Based on the Delaunay triangulation algorithm, any three nodes with edge connections between them Can form a patch , build a complete facet set ; Graph-based node collection , edge set and patch collections , construct a structural diagram of the connected defect area .

[0020] Furthermore, the residual strength of the paperboard material under the microporous defect is predicted The steps include: Geometric feature triples based on connected defect regions , calculate the weighted porosity of paperboard material under microporous defects ; Based on the weighted porosity of the paperboard material , calculate the strength attenuation factor of paperboard material under microporous defects ; Strength attenuation factor of paperboard material based on microporous defects , predicting the residual strength of paperboard materials under microporous defects ; The strength attenuation factor of the paperboard material under the microporous defect The porosity of the cardboard material is calculated by Mapped to the strength attenuation factor of the paperboard material under the corresponding microporous defects , and introduce the fitting parameters and , combined with Mohr-Coulomb empirical model, a nonlinear function reflecting the influence of defects on material strength is constructed; the residual strength of the paperboard material under the micro-hole defect is predicted by multiplying the strength attenuation factor of the paperboard material under the micro-hole defect and the initial strength of the paperboard material .

[0021] Further, the step of calculating the weighted porosity of the paperboard material under the micro-hole defect includes: accumulating the physical areas of all connected defect sub-regions in the geometric feature triplets of the connected defect region , and dividing the total detection area , to obtain the unwighted porosity of the paperboard material ; ; The node connectivity in the structure diagram of the connected defect region is introduced , and the physical area of each connected defect sub-region is structurally weighted to obtain the weighted porosity of the paperboard material .

[0022] Further, the micro-hole defect level evaluation and grading control strategy making includes: based on the residual strength of the paperboard material under the micro-hole defect , the micro-hole defect level is divided, and the quality level of the composite corrugated paperboard is constructed ; based on the quality level of the composite corrugated paperboard , the process adjustment scheme is fused by combining the process experience rule base and the machine learning prediction model, the process adjustment suggestion is corrected, and the final process parameter adjustment suggestion set is constructed ; Further, the micro-hole defect level division includes: three strength grading thresholds are preset , and , satisfying the relationship ; if the residual strength of the paperboard material under the micro-hole defect is greater than or equal to , the quality level of the composite corrugated paperboard is divided into , wherein indicates that the material strength of the composite corrugated paperboard is excellent, the defect influence is negligible, and the process parameters can be maintained as they are; if the residual strength of the paperboard material under the micro-hole defect is greater than or equal to​ and less than , the quality grades of composite corrugated paperboard are divided into ,in It means that the material of the composite corrugated cardboard is qualified but there is a slight decrease in strength, which can be compensated by slightly optimizing the process parameters (such as pressure, temperature, glue ratio, etc.); If the residual strength of the cardboard material under microporous defects Greater than or equal to and less than , the quality grades of composite corrugated paperboard are divided into ,in It indicates that the material of the composite corrugated cardboard has a significant decrease in strength, and it is recommended to adjust the key process parameters to improve the structural stability; If the residual strength of the cardboard material under microporous defects Less than , the quality grades of composite corrugated paperboard are divided into ,in This means that the material of the composite corrugated cardboard has serious defects or has exceeded the safe range of use. It is recommended to scrap it or make major optimization adjustments to the process route.

[0023] Furthermore, the final process parameter adjustment suggestion set is constructed The steps include: Quality grades based on composite corrugated board , analyze the current quality grade information and clarify the process adjustment target accordingly: If the quality grade of the composite corrugated cardboard is , which indicates the residual strength of the paperboard material under microporous defects In the excellent range, no adjustment is required, and the system maintains the current process settings; if the quality grade of the composite corrugated cardboard is or , which indicates the residual strength of the paperboard material under microporous defects If the strength is slightly or moderately low, the system needs to adjust the process parameters to restore the strength level; if the quality grade of the composite corrugated cardboard is , which indicates the residual strength of the paperboard material under microporous defects If there is a serious shortage, the system will start the heavy repair process and simultaneously trigger a manual review prompt or pause mechanism to avoid the continuous output of inferior products; According to the quality grade of composite corrugated cardboard , call the process parameter recommendation set under the corresponding level from the preset process experience rule library, and obtain the process parameter recommendation set of the process experience rule library ; Input the currently detected parameters into the trained machine learning model for prediction, and output the recommended set of process parameters predicted by machine learning. ; Calculate the process parameter recommendation set of the process experience rule base Process parameter recommendation set predicted by machine learning The deviation between the two ; Based on preset tolerance thresholds Relative deviation Compare and execute different fusion strategies: If the relative deviation Less than , then it indicates the recommended value of the empirical rule Predicted values ​​with machine learning The relative deviation rate between Within the tolerance threshold, a simple average is used. If the relative deviation Greater than , then it indicates the recommended value of the empirical rule Predicted values ​​with machine learning The relative deviation rate between Large, confidence weighted fusion is used to obtain the fused process parameter recommendation value ; Check the recommended values ​​of the fusion process parameters Whether it falls within the allowed range of equipment control , if the fusion process parameter recommended value Exceeding the equipment control allowable range When the correction operation is performed, the corrected process parameters are obtained. ; Based on the revised process parameters , build the final process parameter adjustment suggestion set ; The equipment control allowable range is obtained by comprehensively setting equipment specifications and process technology standards, quality and safety constraints, historical experience data statistics and expert craftsman experience thresholds.

[0024] A system for detecting microporous structure defects of low-weight, high-strength composite corrugated cardboard is provided, which is used to implement the above-mentioned method for detecting microporous structure defects of low-weight, high-strength composite corrugated cardboard. The system comprises: High-resolution composite corrugated cardboard image construction module, used to collect high-quality image data of the microporous structure in low-weight and high-strength composite corrugated cardboard, and construct a fused high-resolution composite corrugated cardboard face sheet collection ; Composite corrugated cardboard micropore defect recognition module, based on the fused high-resolution composite corrugated cardboard facet collection , identify micropore defects in composite corrugated cardboard and obtain the binary mask image of micropore defects in composite corrugated cardboard ; The composite corrugated paperboard micro-hole defect spatial relationship analysis and structure reconstruction module is based on a composite corrugated paperboard micro-hole defect binary mask image , spatial relationship analysis and structure reconstruction are performed, and a structure diagram of a connected defect region is constructed ; The micro-hole defect mechanical property prediction module is based on the structure diagram of the connected defect region and the geometric feature triplets of the connected defect region , the influence of the micro-hole defect on the mechanical properties of the paperboard material is analyzed, and the residual strength of the paperboard material under the micro-hole defect is predicted ; The micro-hole defect grade evaluation and process adjustment module is based on the residual strength of the paperboard material under the micro-hole defect , micro-hole defect grade evaluation and grading control strategy formulation are performed.

[0025] Compared with the prior art, the beneficial effects of the present application are: A low-grammage high-strength composite corrugated paperboard micro-hole structure defect detection method and system, first, the low-grammage high-strength composite corrugated paperboard sample is standardized pretreated, including cutting to a uniform size, cleaning the surface and cross section and fixing the sample position, to ensure that the sample size is consistent, there is no surface impurity and the attitude is stable during imaging; On this basis, using industrial multi-focus microscopic imaging technology, according to the preset microscope acquisition parameters, the paperboard micro-hole structure image is collected layer by layer at different focal length levels, and a multi-focus composite corrugated paperboard image is obtained; Subsequently, using the maximum gray scale fusion algorithm, the pixel gray scale values of each focal length layer image are taken maximum point by point, and a fused high-resolution composite corrugated paperboard face sheet set is constructed. This step solves the problem of focal point misalignment caused by uneven distribution of micro-holes in the thickness direction of the paperboard, significantly reduces the difficulty of micro-hole edge recognition caused by single-layer image blur or surface fiber impurities shielding, especially in the micro-hole structure region with small aperture, shallow depth or mixed in the fibers, the accurate presentation of the complete boundary can be realized, and the accurate extraction of the number, size and boundary contour of the micro-holes in the subsequent defect identification process is ensured, providing a high-fidelity image basis for material structure integrity analysis and strength calculation.

[0026] Subsequently, based on the fused high-resolution composite corrugated paper board face sheet set, first, the image is preprocessed by gray scale normalization and two-dimensional Gaussian filtering to enhance the image contrast, smooth the background and suppress random noise generated in the imaging process, and a clearer second composite corrugated paper board face sheet set is obtained; then the trained convolutional neural network model is used to infer the second composite corrugated paper board face sheet set, and the probability value of each pixel corresponding to the micro-hole defect is generated, so as to quantify the position and range of the micro-hole defect in the image; finally, the composite corrugated paper board micro-hole defect probability map is threshold segmented, the area with probability value higher than the set threshold is extracted and binary processing is performed, and the composite corrugated paper board micro-hole defect binary mask map is output, realizing the automatic accurate positioning and area division of the micro-hole defect. This step eliminates the interference of uneven illumination, gray shift and microscopic particle noise in the collected image, ensuring the stable image quality of the subsequent defect recognition model input; further, the model inference replaces manual naked eye recognition, eliminating the visual omission and unclear boundary division caused by the small size of micro-holes, insignificant color difference or dense distribution in traditional manual detection; at the same time, through probability map and threshold segmentation, the continuity and integrity of the micro-hole boundary extraction are ensured, especially in the edge fuzzy or adjacent complex structure area.

[0027] Then, based on the composite corrugated paper board micro-hole defect binary mask map, all independent defect connected regions are identified by spatial connectivity clustering analysis, and then the geometric feature quantitative analysis of each connected region is performed to calculate its area, perimeter, equivalent circle diameter and geometric center coordinates, etc. Finally, taking the defect geometric center of the connected defect region as the node in the graph structure, combining the corresponding geometric features, the Delaunay triangulation algorithm is used to construct the structure graph of the connected defect region, realizing the structured modeling of the micro-hole defect. This step realizes the conversion of scattered and various morphological micro-hole defects from pixel-level binary image to graph structure representation with spatial relationship and geometric information, solving the problem of accurately depicting the mutual position relationship and morphological characteristics between defect groups in traditional defect detection.

[0028] Further, based on the connected defect region structure graph and geometric feature triplets, the influence of micro-hole defects on the mechanical properties of paperboard material is analyzed, and the residual strength of paperboard material under micro-hole defects is predicted. First, the unweighted porosity of the paperboard material is calculated by accumulating the physical area of all connected defect sub-regions and combining the total detection area. Then, the weighted porosity reflecting the intensity of defect distribution is obtained by introducing the node connectivity of each defect region in the structure graph as the weight. Subsequently, based on the weighted porosity of the paperboard material and combined with the empirical model, the strength attenuation factor of the paperboard material under micro-hole defects is calculated, and the weakening degree of the defect on the strength of the paperboard material is quantitatively evaluated. Finally, according to the strength attenuation factor of the paperboard material under micro-hole defects and the initial strength of the paperboard material, the residual strength of the paperboard material under micro-hole defects is calculated, and the prediction and evaluation of the carrying capacity of the composite corrugated paperboard in the actual service condition are realized. This step solves the problem that the traditional method cannot effectively associate the defect distribution characteristics with the mechanical properties, especially suitable for online quality monitoring or product sampling scenarios, and early warning of local micro-defects significantly affecting the compressive strength of the material, so as to timely screen out potential failure products and reduce the risk of batch structure damage caused by local damage in stacking and transportation. Finally, based on the residual strength of the paperboard material under micro-hole defects, the mechanical properties of the composite corrugated paperboard are classified into different quality grades, and the composite corrugated paperboard is divided into different quality grades by setting three strength grading thresholds, and the specific influence degree of defects of different degrees on the carrying capacity of the paperboard material is clear. Then, according to the quality grade of the composite corrugated paperboard, the corresponding process parameter adjustment scheme is retrieved from the process experience rule base and the machine learning prediction model, including key process parameters such as shotcrete amount, pressing pressure, drying temperature, adhesive addition ratio and paper machine line speed, and the deviation analysis of the two is carried out to optimize and ensure that the adjustment suggestion takes into account long-term stability and real-time adaptability. Then, the legal verification and necessary correction of the fused process parameter suggestion value are performed to ensure that all the control parameters meet the operating range allowed by the equipment. Finally, the final process parameter adjustment suggestion set is output to guide the production line to dynamically adjust the key process flow according to different defect grades, such as increasing the shotcrete amount and reducing the paper machine line speed in the low-strength grade sample corresponding batch, to improve the fiber bonding strength and prolong the adhesive reaction time. This step effectively suppresses the strength decline of the material caused by micro-hole defects, avoids the flow of inferior products, reduces the frequency of low-strength paperboard batches from the source, and realizes quality closed-loop control and process precise optimization in the production process. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, brief descriptions will be given to the drawings needed to be used in the embodiments or prior art descriptions. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creating any inventive work.

[0030] Figure 1 is a low-grammage high-strength composite corrugated paperboard micro-pore structure defect detection method of the present application. The principle flow chart of the present application is shown in the figure. Figure 2 is a method flow chart for constructing a fused high-resolution composite corrugated paperboard face sheet set in a low-grammage high-strength composite corrugated paperboard micro-pore structure defect detection method of the present application. Figure 3 is a method flow chart for obtaining a composite corrugated paperboard micro-pore defect binary mask map in a low-grammage high-strength composite corrugated paperboard micro-pore structure defect detection method of the present application. Figure 4 is a method flow chart for constructing a structure map of connected defect regions in a low-grammage high-strength composite corrugated paperboard micro-pore structure defect detection method of the present application. Figure 5 is a method flow chart for predicting the residual strength of the paperboard material under the micro-pore defect in a low-grammage high-strength composite corrugated paperboard micro-pore structure defect detection method of the present application. Figure 6 is a method flow chart for micro-pore defect grade evaluation and grading control strategy development in a low-grammage high-strength composite corrugated paperboard micro-pore structure defect detection method of the present application. Figure 7 is a functional module diagram of a low-grammage high-strength composite corrugated paperboard micro-pore structure defect detection system of the present application. DETAILED DESCRIPTION

[0031] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0032] Embodiment 1 Please refer to Figure 1 The present embodiment provides a low-grammage high-strength composite corrugated paperboard micro-pore structure defect detection method, which comprises: Step S1000, collecting high-quality image data of the micro-pore structure in the low-grammage high-strength composite corrugated paperboard, and constructing a fused high-resolution composite corrugated paperboard face sheet set .

[0033] Specifically, this step aims to collect high-quality image data of the micro-pore structure in the low-grammage high-strength composite corrugated paperboard through standardized sample preprocessing and industrial multi-focus microscopic imaging technology, to provide stable and reliable visual information basis for subsequent micro-pore defect analysis and structure design.

[0034] Further, if Figure 2 As shown, step S1000 includes: Step S1100: Prepare a composite corrugated cardboard sample to obtain a standardized composite corrugated cardboard sample. .

[0035] In the specific implementation process, we first obtain the parameter set of the composite corrugated cardboard sample, which is recorded as: .in, (in ) indicates the Key structural and process attribute information for each composite corrugated cardboard sample, including but not limited to: cardboard batch number information, used to identify samples from different production batches; cardboard grammage (unit: grams per square meter), used to measure the quality of cardboard per unit area and affects the thickness and strength of the material; molding process parameters, including hot pressing temperature, pressure intensity, molding time and other key process control indicators; Indicates the total number of composite corrugated cardboard samples that need to be collected currently.

[0036] Then, based on the composite corrugated cardboard sample parameter set , a standardized process pretreatment process is performed on cardboard samples of different batches and different gram weights, which specifically includes but is not limited to the following steps: First, cutting processing, cutting the original composite corrugated cardboard sample according to the predetermined size specifications to ensure that the sample cross-sectional area is representative and the size is suitable for the clamping platform of the microscopic imaging equipment; Second, surface cleaning, using a dust-free cloth and appropriate solvent to clean the sample surface and cross-section to remove dust, fiber debris or other attached impurities to improve image clarity during imaging; Third, fixed clamping, the processed cardboard sample is firmly installed on the platform of the microscopic imaging device to ensure that the posture remains stable during the imaging process to avoid image blur or structural displacement. After the above series of process processing operations (cutting, cleaning, and fixing), a standardized sample of composite corrugated cardboard is obtained. The standardized sample of the composite corrugated board It has the following three advantages: first, the size specifications are consistent and meet the clamping accuracy requirements of the microscopic imaging equipment; second, the surface and cross-section are clean and uncontaminated, without impurities blocking; third, it is fixed and stable to prevent the sample from moving or tilting.

[0037] Finally, a standardized sample of composite corrugated cardboard The processing parameters and status parameters are recorded as structured data to facilitate subsequent traceability and automated management. The specific expression is as follows: Among them, each (in ) represents the standardized sample after pre-processing, including cutting size (length, width and thickness), cleaning status identification (such as cleanliness level) and fixing method and stability index (clamping strength, position deviation, etc.).

[0038] Step S1200, based on the standardized sample of composite corrugated cardboard , configure the relevant parameters of microscope acquisition, complete microscopic imaging acquisition at different focal lengths, and obtain multi-focus composite corrugated cardboard images .

[0039] In the specific implementation process, the core parameters of the microscopic imaging device are first automatically configured, including the starting focus position (i.e. the height of the first focal layer where the microscope lens is aimed at the sample), focal length step spacing (i.e. the fixed distance the lens moves vertically each time) and the number of focal length layers (That is, the composite corrugated cardboard sample is divided into different focus layers in the vertical direction for shooting, and the image data of different depth of field areas are obtained layer by layer), the total depth of field range is (i.e., the physical distance from the first image focus to the last image focus) and Layer focal height (That is, the lens moves vertically from the initial position The depth of focus moves in steps, indicating the The specific process formula of the focus position when the image is taken is as follows: in, Indicates the image layer number, which is used to mark the focus layer corresponding to each image.

[0040] After the above core parameters are configured, based on each Standardized samples of composite corrugated cardboard at each set focal height The specific process formula is as follows: in, Indicates the Standardized samples of composite corrugated cardboard At focal height Images collected at Represents the image acquisition operation, which is used to acquire images at the specified sample and focal height; The index of the standardized sample image of the composite corrugated cardboard, that is, the index of the composite corrugated cardboard image currently collected; represents the image layer number, i.e. the focal layer corresponding to the current collected composite corrugated paper image.

[0041] Finally, based on each composite corrugated paper image, its corresponding multi-focus composite corrugated paper image is constructed, denoted as . Wherein each represents the normalized sample of the ; thcomposite corrugated paper a set of multi-focus face pieces, represents the normalized sample of the thcomposite corrugated paper in the focal layer corresponding to the thcomposite corrugated paper image.

[0042] For example, taking the shooting of a composite corrugated paper sample as an example, the starting focal position , focal length step distance and focal layer are set, then the first layer focal height , the second layer focal height , and so on. The total depth range is , i.e. the microscope lens collects 6 images with different focal layers from the starting focal position in the vertical direction, fully covering the vertical depth area of the micro-porous structure of the composite corrugated paper in the thickness direction . At the same time, the multi-focus composite corrugated paper image , the image focuses on , the image focuses on , and so on. In the above multi-focus composite corrugated paper image, each image focuses on the micro-porous structure of the paper sample at a certain focal depth.

[0043] Step S1300, based on the multi-focus composite corrugated paper image , a maximum gray scale fusion algorithm is used to construct the fused high-resolution composite corrugated paper face piece set .

[0044] In the specific implementation process, in order to obtain an image with optimal clarity and pore structure details, the maximum gray scale fusion algorithm is used to process the multi-focus composite corrugated paper image . The algorithm realizes the fusion of multi-layer focal point information by selecting the maximum gray scale value in each pixel position in the multi-focus layer, and the specific process formula is as follows: ​ wherein, represents the th fused composite corrugated paperboard image; represents the pixel position coordinate in the image, distributed for the horizontal and vertical pixel positions; represents the th composite corrugated paperboard image sample fused high-resolution image in the pixel position of the gray value; represents the maximum value operator; represents the th composite corrugated paperboard standardization sample in the focal point corresponding to the th composite corrugated paperboard image position of the gray value; represents the paperboard sample index number, a total of paperboard samples; represents the focal length layer index number, a total of images of different focal length layers.

[0045] Finally, based on each fused composite corrugated paperboard image, a set of fused high-resolution composite corrugated paperboard patches is constructed.

[0046] Step S2000, based on the set of fused high-resolution composite corrugated paperboard patches , identify the micro-hole defects in the composite corrugated paperboard, and obtain the composite corrugated paperboard micro-hole defect binary mask image .

[0047] Specifically, this step aims to automatically identify the defect area of the set of fused high-resolution composite corrugated paperboard patches constructed in step S1300, and output a binary defect image , which can be used for subsequent statistical analysis and structure evaluation.

[0048] Further, as shown in Figure 3 , step S2000 includes: Step S2100, based on the set of fused high-resolution composite corrugated paperboard patches , obtain a second set of composite corrugated paperboard patches .

[0049] Specifically, this step mainly pre-processes the set of fused high-resolution composite corrugated paperboard patches , aiming to improve image quality and provide more stable and high-quality input for subsequent deep learning models. The preprocessing process includes two sub-processes of gray scale normalization and noise removal filtering.

[0050] In a specific implementation process, first, the fused high-resolution composite corrugated paperboard face sheet set is subjected to gray scale normalization processing, and the image pixel gray scale value thereof is linearly mapped from to , to obtain a first composite corrugated paperboard face sheet set , and the specific process formula is as follows: wherein represents the gray scale value of the normalized paperboard face sheet set at a pixel point , and the pixel gray scale value ranges from ; represents the gray scale value of the fused high-resolution composite corrugated paperboard face sheet set at a pixel point , and the pixel gray scale value ranges from .

[0051] Subsequently, based on the normalized first composite corrugated paperboard face sheet set , a two-dimensional Gaussian filter is used for noise suppression and image smoothing processing to obtain a second composite corrugated paperboard face sheet set , so as to remove high-frequency noise (such as salt and pepper noise, acquisition interference, etc.) in the image while retaining the edge structure of the image as much as possible, and the specific process formula is as follows: wherein represents the gray scale value of the Gaussian filter denoising paperboard face sheet set at a pixel point ; represents a two-dimensional Gaussian kernel function, which is used to perform weighted average on the surrounding neighborhood with the pixel point as the center; represents a convolution operator; represents the standard deviation of the Gaussian kernel, which controls the extension range and smoothing degree of the kernel function; represents a natural exponential function, which is used to realize the Gaussian attenuation characteristic.

[0052] Step S2200, based on the second composite corrugated paperboard face sheet set , a convolutional neural network model U-Net is used to generate a composite corrugated paperboard micro-hole defect probability map .

[0053] In a specific implementation process, the trained convolutional neural network model U-Net is used to perform inference operation on the second composite corrugated paperboard face sheet set to generate a corresponding composite corrugated paperboard defect probability map The probability map assigns a defect probability value to each pixel position in the image while keeping consistent with the resolution of the input image, which quantifies the possibility of the pixel belonging to the micro-hole defect region and provides a probability basis for the subsequent accurate extraction of the defect region. The specific process formula is as follows: wherein, represents the micro-hole defect probability map of the composite corrugated board, each pixel value is the defect probability value at the pixel position , and the value range is , and represent the height and width of the defect probability map respectively; represents the trained convolutional neural network model U-Net, which is composed of an encoder, a decoder and an activation function; represents the parameter set of the network model, including the convolution kernel weight, the bias term, etc., which is obtained by model training stage learning; represents the Sigmoid activation function, which is used to limit the network output to the interval to realize probability normalization; represents the decoder function, which contains deconvolution or upsampling and skip connection structure, and is used to gradually restore the spatial resolution; represents the encoder function, which is composed of multiple layers of convolution and pooling, and extracts multi-scale semantic features of the image.

[0054] Step S2300, based on the micro-hole defect probability map of the composite corrugated board , threshold segmentation and binary defect region extraction are performed to obtain the micro-hole defect binary mask map of the composite corrugated board .

[0055] In the specific implementation process, based on the micro-hole defect probability map of the composite corrugated board generated in step S2200 , a suitable threshold parameter is set to distinguish the defect probability value of each pixel in the micro-hole defect probability map, and binary processing is performed accordingly to obtain the micro-hole defect binary mask map of the composite corrugated board , which provides a basis for subsequent micro-hole defect morphology analysis. The specific process formula is as follows: wherein, represents the micro-hole defect binary mask map of the composite corrugated board, that is, the pixel value 1 in the map represents that the pixel is judged to be a defect region, and the pixel value 0 represents that the pixel belongs to a normal region; represents the probability value of the corresponding pixel in the micro-hole defect probability map of the composite corrugated board; Threshold parameter for defect judgment decision boundary, which can be adjusted and optimized according to experimental data, actual detection requirements and model performance on the validation set, taking into account the sensitivity and accuracy of defect detection.

[0056] Step S3000, based on the composite corrugated paper micro-hole defect binary mask image , spatial relationship analysis and structure reconstruction are carried out to construct the structure graph of connected defect regions .

[0057] Specifically, this step aims to perform spatial relationship analysis and structure reconstruction on the composite corrugated paper micro-hole defect binary mask image extracted in step S2300, and sequentially complete the identification, geometric feature quantization and structure modeling based on the features of the connected regions of the composite corrugated paper micro-hole defects. The output can be used for composite corrugated paper micro-hole defect morphology modeling, virtual material reconstruction and structure reliability evaluation.

[0058] Further, as shown in Figure 4 , step S3000 includes: Step S3100, based on the composite corrugated paper micro-hole defect binary mask image , spatial connectivity clustering analysis of defect regions is carried out to extract the set of all independent connected defect sub-regions .

[0059] Specifically, this step aims to perform spatial connectivity clustering analysis on all "foreground" pixels (i.e. regions with pixel value 1, indicating potential defect regions) in the composite corrugated paper micro-hole defect binary mask image to identify all spatially independent defect regions in the image, i.e. connected components, laying the foundation for subsequent defect region morphology feature extraction and accurate analysis.

[0060] In the specific implementation process, the connectivity judgment standard is 8-neighbor connectivity. That is: if a foreground pixel and its adjacent upper, lower, left, right four basic direction neighborhood pixels, or left upper, right upper, left lower, right lower four diagonal direction neighborhood pixels, any one of the adjacent pixels is also a foreground pixel (pixel value is 1), then the two are regarded as the same connected region.

[0061] After the above analysis, the set of all independent connected defect sub-regions is extracted , and the specific expression formula is as follows: Among them, represents the set of all independent connected defect sub-regions; represents the total number of connected defect sub-regions detected; Indicates the connectivity defect sub-regions, including all connected A set of pixel coordinates, that is ; Indicates the connectivity defect subregion The The image coordinates of the pixels, represents the horizontal axis (column index), Indicates the vertical coordinate (row index); Indicates the The total number of pixels in a connected region, that is, The number of pixels contained in .

[0062] For example, taking a composite corrugated cardboard defect image as an example, assume that the set of all extracted independent connected defect sub-regions is , four connected defect sub-regions are extracted, specifically: connected region Located in the upper left corner of the image, it is an "L" shaped structure; connected area Distributed in the right middle of the image, in the shape of a vertical line segment; connected area constitute a compact square area; connected area Located at the bottom left of the image, it is distributed in a horizontal band.

[0063] Step S3200: Based on the set of all independent connected defect sub-regions , perform quantitative analysis of geometric features and extract the geometric center coordinates of defects, and construct a geometric feature triplet connecting the defect area and defect geometric center set .

[0064] Specifically, this step aims to extract each connected defect sub-region based on the Conduct quantitative analysis of geometric dimensions, calculate its core geometric parameters, including pixel area, boundary perimeter, and equivalent circular aperture, and construct a result-based defect geometric feature set , providing a quantitative basis for subsequent defect identification, severity determination and quality assessment.

[0065] In the specific implementation process, the set of all independent connected defect sub-regions is traversed in turn. Each defect sub-region in , for each defect sub-region The following geometric feature calculations are performed respectively: First, area calculation. Count each connected defect sub-area The total number of all foreground pixels in , get the The area of ​​the connected defect sub-region , the specific expression formula is as follows: in, Indicates the The area of ​​the connected defect sub-region is the area of ​​the defect sub-region calculated in pixels, and the unit is the square of the pixel.

[0066] If the image is processed by spatial calibration, the The area of ​​the connected defect sub-region Convert to The actual physical area of ​​the connected defect sub-region , in order to conduct a dimensional assessment that is consistent with the actual standard. The specific process formula is as follows: in, Indicates the The physical area of ​​the connected defect sub-region, that is, the area after conversion to the actual physical size, in millimeters squared; Indicates the pixel physical size conversion factor, that is, the side length represented by each pixel in the actual physical size, in millimeters / pixel.

[0067] Second, perimeter extraction. Use boundary extraction algorithms (such as the findContours method in OpenCV) to extract each connected defect sub-region. The boundary pixel set , and then calculate the The perimeter of the connected defect sub-region , the specific process formula is as follows: in, Indicates the connectivity defect subregion The set of boundary points of ; Indicates the connectivity defect subregion The number of boundary points; Represents a set of boundary points The The coordinates of the boundary pixels, represents the horizontal axis (column index), Indicates the vertical coordinate (row index); Indicates the The estimated perimeter of the connected defect sub-region, in pixels; represents the Euclidean distance from the boundary point to .

[0068] Thirdly, the equivalent circle diameter calculation. Approximate each defect region as an ideal circular region with equal area, whose equivalent circle diameter is calculated as follows: Fourthly, the geometric center coordinate extraction. Based on the foreground pixel points in each connected defect sub-region , the geometric center coordinate is extracted, which is calculated as follows: wherein, represents the geometric center coordinate of the th connected defect sub-region ; represents the number of pixel points in the th connected defect sub-region ; represents the index of the pixel point of the th connected defect sub-region , ranging from ; represents the horizontal coordinate (column index) of the th pixel point of the th connected defect sub-region ; represents the vertical coordinate (row index) of the th pixel point of the th connected defect sub-region .

[0069] Finally, based on the above geometric features, a geometric feature triplet of a connected defect region and a set of defect geometric centers are constructed, which are calculated as follows: wherein, represents the physical area of the th connected defect sub-region; represents the perimeter of the th connected defect sub-region; represents the equivalent circle diameter of the th connected defect sub-region; represents the geometric center coordinate of the th connected defect sub-region ; Total number of detected connected defect sub-regions.

[0070] Step S3300, constructing a structure graph of the connected defect regions based on the geometric features of the connected defect regions and the set of defect geometric centers thereof . .

[0071] Specifically, this step aims to establish a structure graph that can quantitatively reflect the spatial layout relationship and topological structure of defects based on the geometric features of the connected defect regions and the geometric feature information of each defect region in the set of defect geometric centers thereof .

[0072] In the specific implementation process, the construction of the structure graph of the connected defect regions includes the following three steps: First, the construction of the graph node set. First, based on the geometric center coordinates of each connected defect sub-region , the geometric center coordinates are regarded as the nodes of the structure graph, and each graph node is bound to the corresponding geometric feature triple as the node attribute, forming the graph nodes , and finally, the complete graph node set is constructed.

[0073] Second, the construction of the graph edge set. First, based on the geometric center coordinates in the graph node set , the connection relationship is established using the Delaunay triangulation algorithm. If the center points of any two nodes , are adjacent in the Delaunay mesh, an undirected edge is established. Finally, the complete graph edge set is constructed, wherein .

[0074] Third, the construction of the face set. Based on the Delaunay triangulation algorithm, any three nodes that are connected by edges can form a face. Finally, the complete face set is constructed. Wherein, represents the th face in the face set, which is an ordered triple consisting of three nodes .

[0075] ​​​​​Finally, based on the node set, edge set and face set of the above graph, a structural graph of the connected defect area is constructed. , the specific expression formula is as follows: Step S4000: Based on the structural diagram of the connected defect area and the geometric feature triples of connected defect regions , analyze the effect of microporous defects on the mechanical properties of cardboard materials, and predict the residual strength of cardboard materials under microporous defects .

[0076] Specifically, this step is to construct the structure diagram of the connected defect area in step S3300. And the geometric feature triplet of the connected defect area constructed in step S3200 , a mechanical properties prediction model for structural sensitivity analysis was established. By introducing the weighted porosity of the paperboard material under microporous defects , Strength attenuation factor of paperboard material under microporous defects and the residual strength of paperboard materials under microporous defects etc., to quantitatively evaluate the residual bearing capacity of composite corrugated cardboard materials in the presence of microporous defects, and thus realize the structured and intelligent judgment of the impact of defects on the mechanical properties of materials.

[0077] Furthermore, if Figure 5 As shown, step S4000 includes: Step S4100: Based on the geometric feature triples of the connected defect area , calculate the weighted porosity of paperboard material under microporous defects .

[0078] Specifically, this step aims to evaluate the proportion of microporous defect areas in the overall material as a basic input parameter for mechanical property prediction.

[0079] In the specific implementation process, first, the geometric feature triples of the connected defect area are The physical area of ​​all connected defect sub-regions in Add the total and divide it by the total detection area , and the unweighted porosity of the paperboard material is obtained , the specific process formula is as follows: in, Represents the unweighted porosity of the cardboard material, that is, the proportion of the defect area in the overall detection area. The value range is ; Indicates the physical area of the connected defect sub-region; total number of detected connected defect sub-regions; total detection area.

[0080] Further, to enhance the responsiveness of the paperboard material structure characteristics to the porosity calculation, a structure graph of the connected defect region is introduced node connectivity , for each structure weighting to obtain the weighted porosity of the paperboard material The specific process formula is as follows: wherein, weighted porosity of the paperboard material, used to reflect the weighted influence of defects on material performance; structure weighting factor, used to the first defect sub-region importance weight of performance influence in the structure graph; structure weight adjustment parameter, which needs to be obtained through experimental optimization (usually positive); node connectivity connectivity of the first defect sub-region in the graph structure, used to count the number of adjacent edges of each node , i.e. the connection activity of the structure graph, if the boundary distance between any two defect sub-regions and is less than a certain threshold, there is an adjacent edge number .

[0081] Step S4200, based on the weighted porosity of the paperboard material , calculate the strength attenuation factor of the paperboard material under the micro-hole defect. .

[0082] Specifically, this step aims to further evaluate the degree of weakening of the micro-hole defect to the strength performance of the paperboard material based on the weighted porosity of the paperboard material calculated in step S4100 .

[0083] In the specific implementation process, the weighted porosity of the paperboard material is mapped to the strength attenuation factor of the paperboard material under the micro-hole defect , fitting parameters and are introduced, and a nonlinear function relationship of the influence of defects on strength is constructed combining the Mohr-Coulomb empirical model to calculate the strength attenuation factor of the paperboard material under the micro-hole defect , for quantifying the degree of weakening of the overall strength of the paperboard material by the micro-hole defect, the specific expression formula is as follows: , wherein, represents the strength attenuation factor of the paperboard material under the micro-hole defect, for representing the weakening proportion of the strength of the paperboard material by the defect, and the smaller the value is, the more the strength decreases; represents an empirical attenuation coefficient, which is usually obtained through experiments or literature fitting; represents a nonlinear index parameter, for controlling the curvature of the nonlinear attenuation, which is usually greater than 1, enhancing the sensitivity to high porosity, and is usually obtained through experiments or literature fitting.

[0084] Step S4300, predicting the residual strength of the paperboard material under the micro-hole defect based on the strength attenuation factor of the paperboard material under the micro-hole defect . .

[0085] Specifically, the present step aims to calculate the effective residual carrying capacity of the paperboard material under the current defect condition, based on the strength attenuation factor of the paperboard material under the micro-hole defect calculated in step S4200 , in combination with the theoretical initial strength of the paperboard material in the ideal defect-free state.

[0086] In the specific implementation process, in combination with the strength attenuation factor of the paperboard material under the micro-hole defect and the initial strength of the paperboard material , the residual strength of the paperboard material under the micro-hole defect , i.e. the effective carrying capacity of the paperboard material under the current defect condition, is predicted, and the specific process formula is as follows: , wherein, represents the residual strength of the paperboard material under the micro-hole defect, for the effective carrying capacity of the material under the defect influence; represents the initial strength of the material, i.e. the designed theoretical strength of the material when it is not affected by the defect, which can be obtained through standard mechanical tests (such as compression test, tensile test and three-point bending test, etc.).

[0087] Step S5000, performing micro-hole defect grade evaluation and grading control strategy formulation based on the residual strength of the paperboard material under the micro-hole defect .

[0088] Specifically, the present step aims to calculate the effective residual carrying capacity of the paperboard material under the current defect condition, based on the residual strength of the paperboard material under the micro-hole defect calculated in step S4300 The degree of influence of the micro-hole defect is graded, and a corresponding process adjustment suggestion is generated accordingly, so as to realize the quality closed-loop control in the production process of the low-grammage high-strength corrugated paperboard.

[0089] Further, as shown in Figure 6 , step S5000 includes: Step S5100, based on the residual strength of the paperboard material under the micro-hole defect , grading the micro-hole defect, constructing the quality grade of the corrugated paperboard.

[0090] Specifically, the present step aims to grade the mechanical properties of the current corrugated paperboard based on the residual strength of the paperboard material under the micro-hole defect calculated in step S4300, so as to realize the grading management of the defect evaluation result and the process adjustment reference.

[0091] In the specific implementation process, three strength grading thresholds , and are preset, which satisfy the relationship and the grading thresholds can be obtained by training sample data fitting according to the standard or production experience, for interval division of the residual strength of the paperboard material under the micro-hole defect . Based on the strength grading thresholds , and , the quality grading rules of the corrugated paperboard are constructed, and the specific process formula is as follows: If , the quality grade of the corrugated paperboard is . Among them, indicates that the material strength of the corrugated paperboard is excellent, the defect influence is negligible, and the process parameters can be maintained as they are; If , the quality grade of the corrugated paperboard is . Among them, indicates that the material of the corrugated paperboard is qualified but there is a slight strength decrease, which can be compensated by small amplitude optimization of process parameters (such as pressure, temperature, glue ratio, etc.); If , the quality grade of the corrugated paperboard is . Among them, indicates that the material of the corrugated paperboard has obvious strength decrease, and it is suggested to adjust the key process parameters to improve the structural stability; If , the quality grade of the corrugated paperboard is . Among them, ​This means that the material of the composite corrugated cardboard has serious defects or has exceeded the safe range of use. It is recommended to scrap it or make major optimization adjustments to the process route.

[0092] Ultimately, the quality grade of composite corrugated board .

[0093] Step S5200, based on the quality grade of the composite corrugated cardboard , integrate the process adjustment plan of the process experience rule library and the machine learning prediction model, revise the process adjustment suggestions, and build the final process parameter adjustment suggestion set .

[0094] In the specific implementation process, according to the quality grade of the composite corrugated cardboard The matching process adjustment plan is retrieved from the process experience rule library and the trained machine learning prediction model, and a set of process parameter adjustment suggestions containing multiple control dimensions is automatically generated. The process parameter adjustment suggestion set It covers multiple key control parameters such as the spraying system, pressure system, drying system, adhesive dosing system and paper machine line speed, in order to suppress defects in composite corrugated paperboard and improve the residual strength of the paperboard. The specific process is as follows: First, based on the quality grade of composite corrugated cardboard , analyze the current quality grade information and clarify the process adjustment target accordingly: If the quality grade of the composite corrugated cardboard is , which indicates the residual strength of the paperboard material under microporous defects In the excellent range, no adjustment is required, and the system maintains the current process settings; if the quality grade of the composite corrugated cardboard is or , which indicates the residual strength of the paperboard material under microporous defects If the strength is slightly or moderately low, the system needs to adjust the process parameters to restore the strength level; if the quality grade of the composite corrugated cardboard is , which indicates the residual strength of the paperboard material under microporous defects If there is a serious shortage, the system will start a heavy repair process and simultaneously trigger a manual review prompt or pause mechanism to avoid the continuous output of inferior products.

[0095] Second, according to the quality grade of composite corrugated cardboard , call the process parameter recommendation set under the corresponding level from the preset process experience rule library, and obtain the process parameter recommendation set of the process experience rule library , covering but not limited to the following dimensions: shotcrete volume (unit: ), pressing pressure (Unit: MPa), drying temperature (unit: ), adhesive dosage ratio (unit: ) and paper machine line speed The empirical values ​​and adjustment ranges of parameters such as Indicates the The empirical rule recommended values ​​of the process parameters; Indicates the rule suggestion scenario; Represents the total number of process parameters. The rule base is constructed based on historical production line experience, process engineer recommendations, and statistical analysis results to form a robust and reliable parameter adjustment range. For example, if the current quality grade of the composite corrugated cardboard is , the system can recommend increasing the amount of spraying , drying temperature increases , and slow down the paper machine line speed by 2%.

[0096] Third, to further enhance the personalization and adaptability of parameter recommendations, the system inputs the currently detected physical parameters (such as cardboard moisture, thickness, initial strength), raw material information, and production environment parameters into a trained machine learning model for prediction. The model can be a multi-layer perceptron neural network (MLP), random forest (RF), or support vector machine (SVM) structure, and its training data comes from historical samples of correlation between cardboard strength and process parameters. The prediction model outputs a set of process parameter recommendations predicted by machine learning. , as a supplement or correction to the empirical rules, to improve the dynamic response capability of the suggestions. Indicates the Machine learning predictions of process parameters; Indicates the machine scenario; Indicates the total number of process parameters; Fourth, the process parameter recommendation set of the process experience rule library Process parameter recommendation set predicted by machine learning Fusion is performed to generate the final set of process parameter adjustment recommendations First, calculate the process parameter recommendation set of the process experience rule base Process parameter recommendation set predicted by machine learning The deviation between the two The specific process formula is as follows: in, Indicates the The relative deviation rate of the process parameters is used to measure the closeness between the two (value range ); Indicates the empirical rule recommended value of the process parameter; machine learning predicted value of the process parameter; empirical rule recommended value of the process parameter; the following fusion strategy is performed based on a preset tolerance threshold: if , it indicates that the relative deviation rate between the empirical rule recommended value and the machine learning predicted value is within the tolerance threshold range, and a simple average value is adopted, and the specific process formula is as follows: if , it indicates that the relative deviation rate between the empirical rule recommended value and the machine learning predicted value is greater, and a confidence weighted fusion is adopted, and the specific process formula is as follows: wherein, represents the weight of the empirical rule recommended value; represents the weight of the machine learning predicted value; represents the success rate of the process parameter suggestion set of the process empirical rule library based on historical data statistics, i.e. , represents the actual application times of the process parameter suggestion set of the process empirical rule library, represents the times of reaching the expected effect in the process parameter suggestion set of the process empirical rule library; represents the confidence index of the prediction result of the process parameter in the machine learning predicted process parameter suggestion set, which is used to measure the credibility of the prediction value, and is usually directly output or indirectly calculated by the machine learning model; represents the fused process parameter suggestion value.

[0097] Subsequently, parameter legality verification and correction. It is checked whether the fused process parameter suggestion value falls within the equipment control allowable interval . The equipment control allowable interval is obtained by comprehensively setting the equipment specifications and process technical standards, quality safety constraints, historical experience data statistics and expert process threshold.

[0098] if the fused process parameter suggestion value exceeds the equipment control allowable interval , the following correction operation needs to be performed: ​​​​ If , then set to lower limit ; if , then set to upper limit ; if , no correction is needed.

[0099] For example, assuming that the control interval of the pressing pressure of a certain sample of composite corrugated paperboard is , the system-fused process parameter recommended value is , the correction operation is performed, and the corrected process parameter is finally obtained.

[0100] Finally, based on the corrected process parameter , the final process parameter adjustment recommendation set is constructed, which is used as the basis for automatic or manual process adjustment recommended to the site to guide the control of key links such as spraying, pressing, drying, bonding, and running speed, so as to realize the stable recovery and quality improvement of paperboard strength.

[0101] Embodiment 2 Based on Embodiment 1, this embodiment provides a low-grammage high-strength composite corrugated paperboard micro-hole structure defect detection system, as shown in Figure 7 , which comprises: a high-resolution composite corrugated paperboard image construction module, configured to collect high-quality image data of the micro-hole structure in the low-grammage high-strength composite corrugated paperboard, and construct a fused high-resolution composite corrugated paperboard face sheet set ; a composite corrugated paperboard micro-hole defect recognition module, configured to recognize micro-hole defects in the composite corrugated paperboard based on the fused high-resolution composite corrugated paperboard face sheet set , and obtain a composite corrugated paperboard micro-hole defect binary mask image ; a composite corrugated paperboard micro-hole defect spatial relationship analysis and structure reconstruction module, configured to perform spatial relationship analysis and structure reconstruction based on the composite corrugated paperboard micro-hole defect binary mask image , and construct a structure graph of connected defect regions ; a micro-hole defect mechanical property prediction module, configured to analyze the influence of the micro-hole defect on the mechanical properties of the paperboard material based on the structure graph of the connected defect regions and the geometric feature triplets of the connected defect regions , and predict the residual strength of the paperboard material under the micro-hole defect ; a micro-hole defect grade evaluation and process adjustment module, configured to evaluate the grade of the micro-hole defect based on the residual strength of the paperboard material under the micro-hole defect​ The micro-hole defect level is evaluated and a grading control strategy is made.

[0102] The part of the above technical solution provided in the embodiments of the present application which is consistent with the implementation principle of the corresponding technical solution in the prior art is not described in detail to avoid excessive description.

[0103] The specific embodiments described above further illustrate the objects, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for detecting microporous structural defects in low-weight and high-strength composite corrugated paperboard, characterized in that: include: Collect high-quality image data of the microporous structure of low-weight, high-strength composite corrugated cardboard and construct a fused high-resolution composite corrugated cardboard image set; Based on the fused high-resolution composite corrugated cardboard image set, the micropore defects in the composite corrugated cardboard are identified and a binary mask image of the micropore defects in the composite corrugated cardboard is obtained; Based on the binary mask image of micropore defects in composite corrugated cardboard, spatial relationship analysis and structural reconstruction are performed to construct a structural map of connected defect areas. Based on the structural diagram of the connected defect area and the geometric feature triplet of the connected defect area, the influence of microporous defects on the mechanical properties of paperboard materials is analyzed, and the residual strength of paperboard materials under microporous defects is predicted. Based on the residual strength of the paperboard material under microporous defects, microporous defect grade evaluation and graded control strategy are formulated.

2. The method for detecting microporous structure defects of low-weight and high-strength composite corrugated cardboard according to claim 1, characterized in that: The step of constructing a fused high-resolution composite corrugated cardboard image set comprises: Obtaining a standardized sample of composite corrugated cardboard and obtaining a multi-focus composite corrugated cardboard image; Based on the multi-focus composite corrugated cardboard images, the maximum grayscale fusion algorithm is used to construct a fused high-resolution composite corrugated cardboard image set.

3. The method for detecting microporous structure defects of low-weight and high-strength composite corrugated paperboard according to claim 1, characterized in that: The step of obtaining a binary mask image of micropore defects of the composite corrugated paperboard comprises: Based on the fused high-resolution composite corrugated cardboard image set, a second composite corrugated cardboard image set is obtained; Based on the second composite corrugated cardboard image set, a convolutional neural network model is used to generate a composite corrugated cardboard micropore defect probability map; Based on the micropore defect probability map of the composite corrugated cardboard, threshold segmentation and binary defect area extraction are performed to obtain the binary mask map of the micropore defect of the composite corrugated cardboard.

4. The method for detecting microporous structure defects of low-weight and high-strength composite corrugated paperboard according to claim 1, characterized in that: The steps of spatial relationship analysis and structure reconstruction include: Based on the binary mask image of micropore defects in composite corrugated cardboard, a cluster analysis of the spatial connectivity of the defect area is performed to extract the set of all independent connected defect sub-areas. Based on the set of all independent connected defect sub-regions, quantitative analysis of geometric features and extraction of defect geometric center coordinates are performed to construct the geometric feature triples of connected defect regions and the set of defect geometric centers. Based on the geometric features of the connected defect region and the set of defect geometric centers, a structural graph of the connected defect region is constructed.

5. The method for detecting microporous structure defects of low-weight and high-strength composite corrugated paperboard according to claim 4, characterized in that: The spatial connectivity cluster analysis of the defect area is to use the 8-neighborhood connectivity rule to perform connectivity judgment on all foreground pixels in the binary mask image of the micropore defects of the composite corrugated cardboard; The 8-neighborhood connectivity judgment specifically includes: if a foreground pixel and any of its adjacent neighboring pixels in the four basic directions of up, down, left, and right, or any of its adjacent pixels in the four diagonal directions of upper left, upper right, lower left, and lower right are also foreground pixels, then the two are considered to be the same connected area.

6. The method for detecting microporous structure defects of low-weight and high-strength composite corrugated paperboard according to claim 4, characterized in that: The steps of calculating the geometric features include: Count the total number of foreground pixels in each connected defect sub-region, calculate the area of ​​each connected defect sub-region, and after the image is spatially calibrated, convert the area of ​​each connected defect sub-region to obtain the actual physical area of ​​each connected defect sub-region; A boundary extraction algorithm is used to extract the boundary pixel points of each connected defect sub-region, and the perimeter of each connected defect sub-region is calculated based on the accumulation of Euclidean distances between adjacent boundary points. Each defect area is approximated as an ideal circular area of ​​equal area, and the equivalent circle diameter of each connected defect sub-area is calculated accordingly; Based on the foreground pixels in each connected defect sub-region, the average values ​​of the horizontal and vertical coordinates of each foreground pixel are counted respectively, and the geometric center coordinates of each connected defect sub-region are extracted.

7. The method for detecting microporous structure defects of low-weight and high-strength composite corrugated paperboard according to claim 1, characterized in that: The step of constructing a structural diagram of connected defect areas comprises: Based on the geometric center coordinates of each connected defect sub-region, it is regarded as a node of the structure graph, and each graph node is bound to the corresponding geometric feature triple as the node attribute to form a graph node and construct a complete graph node set; Based on the coordinates of the geometric centers of the nodes in the graph, the Delaunay triangulation algorithm is used to establish the connection relationship. If the centers of any two nodes are adjacent in the Delaunay grid, an undirected edge is established to construct a complete graph edge set. Based on the Delaunay triangulation algorithm, any three nodes connected by edges can form a face, thus constructing a complete face set; Based on the node set, edge set and face set of the graph, a structural graph of the connected defect area is constructed.

8. The method for detecting microporous structure defects of low-weight and high-strength composite corrugated paperboard according to claim 1, characterized in that: The step of predicting the residual strength of the paperboard material under microporous defects includes: The weighted porosity of paperboard material with microporous defects is calculated based on the geometric characteristic triples of connected defect regions. Based on the weighted porosity of the cardboard material, the strength attenuation factor of the cardboard material under microporous defects is calculated; Based on the strength attenuation factor of the paperboard material with microporous defects, the residual strength of the paperboard material with microporous defects is predicted.

9. A method for detecting microporous structural defects of low-weight and high-strength composite corrugated paperboard according to claim 8, characterized in that: The step of calculating the weighted porosity of the paperboard material under microporous defects includes: accumulating the physical areas of all connected defect sub-regions in the geometric feature triples of the connected defect region and dividing the sum by the total detection area to obtain the unweighted porosity of the paperboard material; introducing the node connectivity in the structural graph of the connected defect region and structurally weighting the physical area of ​​each connected defect sub-region to obtain the weighted porosity of the paperboard material.

10. A system for detecting microporous structure defects of low-weight and high-strength composite corrugated cardboard is used to implement a method for detecting microporous structure defects of low-weight and high-strength composite corrugated cardboard according to any one of claims 1 to 9, characterized in that: The system comprises: High-resolution composite corrugated cardboard image construction module, used to collect high-quality image data of the microporous structure in low-weight, high-strength composite corrugated cardboard, and construct a fused high-resolution composite corrugated cardboard image set; The composite corrugated cardboard micropore defect recognition module identifies micropore defects in the composite corrugated cardboard based on the fused high-resolution composite corrugated cardboard image set and obtains a binary mask image of the composite corrugated cardboard micropore defects; The module for spatial relationship analysis and structural reconstruction of micropore defects in composite corrugated cardboard performs spatial relationship analysis and structural reconstruction based on the binary mask map of micropore defects in composite corrugated cardboard, and constructs a structural map of connected defect areas. The microporous defect mechanical properties prediction module analyzes the impact of microporous defects on the mechanical properties of paperboard materials based on the structural diagram of the connected defect area and the geometric feature triples of the connected defect area, and predicts the residual strength of the paperboard material under microporous defects; The micropore defect grade assessment and process adjustment module assesses the micropore defect grade and formulates a graded control strategy based on the residual strength of the paperboard material under micropore defects.

Images