An image processing-based real-time monitoring method for production of chlorine-dissociating anode plates

Abstract

The application discloses a kind of based on image processing's real-time monitoring method of chlor-alkali type anode plate production, it is related to image processing technical field, comprising the following steps: S1, by thermal field modeling is pre-calibrated;S2, build coating pyrolysis luminescence spectrum model, image is reconstructed in space;S3, build defect evolution model, and judge whether there is potential defect risk;S4, by irregular clustering mechanism clustering, form abnormal area map;S5, establish pyrolysis defect intervention control model, guide real-time process adjustment;S6, build time period stability scoring model, carries out trend evaluation.The application introduces thermal field modeling calibration mechanism in pyrolysis image acquisition stage, improves the perception sensitivity of local fluctuation of pyrolysis behavior;By building luminescence spectrum model and multi-period defect evolution modeling, realizes the fine monitoring and dynamic evolution judgment to defect generation process;Irregular clustering method breaks through the partition limitation of traditional fixed template.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to a real-time monitoring method for the production of chlorine-evolving anode plates based on image processing. Background Technology

[0002] Chlorine-evolving anode plates are widely used in the chlor-alkali industry, and their quality directly affects electrolysis efficiency and service life. In existing production processes, the pyrolysis stage is prone to defects such as cracks, warping, and residual bubbles due to high temperatures, opacity, and frequent physical disturbances. Traditional quality inspection methods rely heavily on manual sampling, infrared thermography, or static image analysis, which are not only slow to respond but also lack effective tracking of abnormal evolution patterns during pyrolysis, making real-time monitoring and intelligent intervention of the entire process difficult.

[0003] Especially in the field of image processing, existing methods generally rely on single-scale feature extraction with fixed thresholds, which cannot adapt to complex surface morphology changes, nor can they establish multi-stage, multi-scale defect development models, resulting in the inability to predict quality risks in advance. At the same time, most current defect clustering and partitioning methods are statically set and lack modeling support for the spatiotemporal evolution of defect regions, which seriously restricts the accuracy of anomaly warning and the precision of process control. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a real-time monitoring method for the production of chlorine-evolving anode plates based on image processing, thereby solving the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] In a first aspect, embodiments of the present invention provide a real-time monitoring method for the production of chlorine-evolving anode plates based on image processing, comprising the following steps:

[0007] S1. Construct a pyrolysis window image acquisition system and pre-calibrate the acquired images through thermal field modeling;

[0008] S2. Construct a coating pyrolysis luminescence spectrum model and spatially reconstruct the image calibrated in step S1.

[0009] S3. Based on the data output in step S2, construct a defect evolution model and determine whether there are potential defect risks.

[0010] S4. Based on the results output by the defect evolution model, clustering is performed using an irregular clustering mechanism to form an anomaly region map;

[0011] S5. Based on the results of the irregular clustering mechanism, a pyrolysis defect intervention and control model is established to guide real-time process adjustment;

[0012] S6. Construct a time-period stability scoring model to evaluate the trend of the anode plate production process under real-time monitoring.

[0013] To further optimize this technical solution, in step S1, the pyrolysis window image acquisition system includes an infrared quartz observation window and an industrial-grade high-resolution infrared camera arranged outside it, used to acquire the luminescence image of the surface of the coated anode plate titanium substrate during the pyrolysis process.

[0014] To establish a quantifiable mapping relationship between luminescence intensity and the real temperature field, the following thermal field calibration model is designed through thermal field modeling:

[0015]

[0016] in,

[0017] : The intensity of the fitted thermal field corresponding to pixel (x, y);

[0018] The grayscale value of a pixel in the infrared channel;

[0019] Background reference grayscale base;

[0020] All of these are calibration constants obtained by pre-fitting the system through a standard blackbody temperature field experiment;

[0021] The thermal field calibration model is used to ensure that the intensity values ​​of each subsequent pixel have thermophysical meaning in spatial coordinates, thereby constructing the change and evolution structure of the thermal decomposition spectrum.

[0022] To further optimize this technical solution, in step S2, an image with thermophysical significance is obtained using the thermal field calibration model established in step S1. Based on this image, a spectral structure is constructed, and the output from step S1 is... Transformed into time series It also enables spatial scene reconstruction, providing input data structures for subsequent defect identification;

[0023] The coating surface is divided into multiple grid sub-regions, and a coating pyrolysis luminescence spectrum model is established for the time evolution of each region:

[0024]

[0025] in,

[0026] : The average luminous intensity of the i-th and j-th sub-regions at time t;

[0027] : The pixel size of each grid sub-block;

[0028] : This refers to the data output in step S1 after the transformation;

[0029] The coating pyrolysis luminescence spectrum model structures the original image into a regular spatial grid, constructing regional units that are easy to process mathematically; and captures the luminescence change trajectory of the coating at the same location during the pyrolysis process, providing an evolutionary basis for identifying defect points.

[0030] To further optimize this technical solution, in step S3, during the production of the chlorine-evolving anode plate, defects include localized insufficient sintering, coating peeling, and over-sintering. These defects manifest as abnormal luminescence intensity regions during pyrolysis. The defect evolution model is constructed based on the local incremental change rate as follows:

[0031]

[0032] in,

[0033] : The range of the rate of change of pyrolysis intensity in the i-th and j-th subregions;

[0034] Data from the output of step S2;

[0035] When a region meets the following conditions:

[0036]

[0037] That is, the difference in the rate of change exceeds the system's set threshold. The pyrolysis process in this area was determined to be unstable, posing a potential risk of defects.

[0038] By using a defect evolution model, "abnormal evolution trends" can be identified in advance before visible cracks or detachment occur, and early warning signals can be generated as triggering conditions for subsequent closed-loop control.

[0039] To further optimize this technical solution, in step S4, during the pyrolysis of the anode plate, the abnormal area is affected by the uneven micro-coating and the thermal conductivity effect of the substrate, resulting in a "band-like" or "sheet-like" structure.

[0040] By using irregular clustering mechanisms to locate abnormal distribution structures, the abnormal distribution structures in step S3 are... Mapped to a global anomaly region map.

[0041] To further optimize this technical solution, the irregular clustering mechanism is designed with the following clustering sensitivity mapping function:

[0042]

[0043] in,

[0044] : The comprehensive pyrolysis anomaly value of the k-th cluster anomaly region;

[0045] : The set of indexes for all subregions belonging to cluster k;

[0046] Weighting factors are dynamically adjusted based on location, adjacency, and hotspot density.

[0047] Based on comprehensive pyrolysis anomalies The final output is an anomaly region map of the clustering mapping, and the boundaries and intensity of each cluster region in the map are used to guide pyrolysis intervention.

[0048] To further optimize this technical solution, in step S5, the pyrolysis intervention control model uses the output from step S4. The model is shown below as an input parameter:

[0049]

[0050] in,

[0051] : Intervention urgency score for the k-th cluster region;

[0052] Abnormal evolution speed;

[0053] : The pyrolysis initiation delay parameter of the region;

[0054] Weight parameters.

[0055] To further optimize this technical solution, the pyrolysis intervention control model also includes a comparison limit value. ;

[0056] when When this occurs, the corresponding control module is automatically triggered, issuing the following intervention strategy command:

[0057] Lower the local pyrolysis temperature;

[0058] Delaying or halting subsequent sintering in a certain area;

[0059] Increase the airflow velocity in this area;

[0060] Adjust the position of the carrier plate to avoid localized overheating.

[0061] To further optimize this technical solution, in step S6, the time-period stability scoring model is used to analyze the defect control effect of the anode plate in multiple pyrolysis stages:

[0062]

[0063] in,

[0064] : The average defect suppression effect index during the s-th time period;

[0065] : Abnormal values ​​of comprehensive pyrolysis before intervention;

[0066] The overall pyrolysis anomaly value obtained from the reassessment after intervention;

[0067] Avoid using tiny constants with a denominator of zero;

[0068] The total number of abnormal clusters monitored during this period.

[0069] To further optimize this technical solution, the method also performs feature compression on the abnormal region map formed in the pyrolysis image of each batch of anode plate coating and constructs a defect map fingerprint with high-dimensional vector description.

[0070] The graph fingerprints are persistently stored and assigned batch tags, forming a defect memory bank that is "traceable, transferable, and matchable".

[0071] In a second aspect, embodiments of the present invention provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the computer program instructions, when executed by the processor, implement the steps of a real-time monitoring method for the production of a chloride-evolving anode plate based on image processing as described in the first aspect of the present invention.

[0072] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, they implement the steps of a real-time monitoring method for the production of a chloride-evolving anode plate based on image processing as described in the first aspect of the present invention.

[0073] Compared with existing technologies, this invention provides a real-time monitoring method for the production of chlorine-evolving anode plates based on image processing, involving machine learning and deep learning technologies, and has the following beneficial effects:

[0074] This image processing-based real-time monitoring method for chlorine-evolving anode plate production improves the sensitivity to local fluctuations in pyrolysis behavior by introducing a thermal field modeling and calibration mechanism during the pyrolysis image acquisition stage. Through the construction of a luminescence spectrum model and multi-time-period defect evolution modeling, it achieves refined monitoring and dynamic evolution judgment of the defect generation process. The irregular clustering method overcomes the limitations of traditional fixed template partitioning, flexibly adapting to complex surface defect morphologies. The system can accurately quantify the evolution trend of anode plate defects and possesses multiple capabilities including anomaly warning, intervention verification, and process optimization decision-making. Compared with existing technologies, this solution has significant advantages such as strong real-time performance, high predictability, and strong adaptability to complex pyrolysis scenarios. Attached Figure Description

[0075] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0076] Figure 1 This is a flowchart illustrating a real-time monitoring method for the production of chlorine-evolving anode plates based on image processing, as proposed in this invention. Detailed Implementation

[0077] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0078] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0079] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0080] Example 1:

[0081] Reference Figure 1 This is the first embodiment of the present invention, which provides a real-time monitoring method for the production of chlorine-evolving anode plates based on image processing, including the following steps:

[0082] S1. Construct a pyrolysis window image acquisition system and pre-calibrate the acquired images through thermal field modeling.

[0083] The pyrolysis window image acquisition system includes an infrared quartz observation window and an industrial-grade high-resolution infrared camera arranged outside it. By using a custom multi-band filter (near-infrared and visible light composite channel), it acquires the luminescence image of the surface of the coated anode plate titanium substrate during the pyrolysis process.

[0084] Meanwhile, to ensure the alignment accuracy of each image in spatial coordinates, a calibration crosshair is set in front of the viewing window. The system automatically corrects the viewing angle distortion and projection offset during the image preprocessing stage, thereby ensuring strict consistency between images acquired at different times.

[0085] To establish a quantifiable mapping relationship between luminescence intensity and the real temperature field, the following thermal field calibration model is designed through thermal field modeling:

[0086]

[0087] in,

[0088] : The intensity of the fitted thermal field corresponding to pixel (x, y);

[0089] The grayscale value of this point in the infrared channel;

[0090] Background reference grayscale base;

[0091] All of these are calibration constants obtained by pre-fitting the system through a standard blackbody temperature field experiment.

[0092] This model uses a logarithmic transformation of grayscale values ​​in infrared images, combined with calibration parameters obtained beforehand from standard blackbody temperature field experiments. The infrared grayscale value of each pixel (x, y) is... Mapped to field intensity with pyrolysis significance .in As a background reference value, it represents the image baseline when there is no pyrolysis reaction in the initial state of the system.

[0093] The model acquires infrared images of the current anode plate pyrolysis process, iterates through each pixel in the image, calculates the corresponding pyrolysis intensity according to the model, and then... The data is reconstructed into a two-dimensional thermal field distribution map, providing input for subsequent dynamic analysis.

[0094] The thermal field calibration model is used to ensure that the intensity values ​​of each subsequent pixel have thermophysical meaning in spatial coordinates, thereby constructing the change and evolution structure of the thermal decomposition spectrum.

[0095] Unlike traditional methods that rely solely on visible light or single infrared images for coarse recording, this approach constructs a pyrolysis window image acquisition system and simultaneously introduces a thermal field calibration mechanism, enabling each pixel in the image to possess true thermophysical meaning. The greatest advantage of this method is its ability to convert "image grayscale values" into "pyrolysis intensity values," achieving quantitative spatial thermal field reconstruction rather than merely serving as a "visual reference" for defects. Furthermore, the introduction of an infrared quartz observation window in conjunction with a custom filter significantly improves the imaging contrast for weak oxidation luminescence processes, thus meeting the visualization requirements of the dynamic changes of chlorine-evolving anode plates at high temperatures.

[0096] S2. Construct a coating pyrolysis luminescence spectrum model and spatially reconstruct the image calibrated in step S1.

[0097] Using the thermal field calibration model established in step S1, an image with thermophysical significance has been obtained. Based on the image, a spectral structure is constructed, and the output from step S1 is... Transformed into time series It also enables spatial scene reconstruction, providing input data structures for subsequent defect identification;

[0098] The coating surface is divided into multiple grid sub-regions, and a coating pyrolysis luminescence spectrum model is established for the time evolution of each region:

[0099]

[0100] in,

[0101] : The average luminous intensity of the i-th and j-th sub-regions at time t, expressed in optical grayscale values;

[0102] : The pixel size of each grid sub-block, i.e., the width and height of the grid (in pixels). It is an average of the number of pixels and is a dimensionless normalization coefficient;

[0103] : This is the data output in step S1 after conversion. It has thermophysical meaning and expresses the pyrolysis luminescence intensity of the chlorine-evolving anode plate at position (x, y) as a function of time t. It is used to describe the spatiotemporal characteristics of defect evolution and is expressed in optical grayscale value.

[0104] The original image is divided into a fixed grid, and the pyrolysis intensity of all pixels within each small region is averaged to obtain the luminescence level of that region at the current time point. For each grid region, a luminescence intensity variation curve is constructed over time to form a dynamic spectral data structure.

[0105] The coating pyrolysis luminescence spectrum model structures the original image into a regular spatial grid, constructing regional units that are easy to process mathematically; and captures the luminescence change trajectory of the coating at the same location during the pyrolysis process, providing an evolutionary basis for identifying defect points.

[0106] Current thermal monitoring technologies for chlorine-evolving anode plates mostly rely on static anomaly identification based on single-frame images, lacking in-depth analysis of the temporal evolution of the images, particularly in dynamically monitoring the pyrolysis state of the coating and the evolution trend of anomalous regions. This invention, however, constructs a coating pyrolysis luminescence spectrum model, structuring image information into a time-series spatial grid pyrolysis data volume, and achieving a traceable and reconstructable spectrum structure. The advantages of this method are: it not only provides regional-level pyrolysis intensity representation but also establishes a trajectory model of pyrolysis changes over time, making the generation, diffusion, and aggregation processes of anomalous regions predictable and analyzable, significantly improving the real-time nature of monitoring and the forward-looking nature of diagnosis.

[0107] S3. Based on the data output in step S2, construct a defect evolution model and determine whether there are potential defect risks.

[0108] In the production process of chlorine-evolving anode plates, defects include localized insufficient sintering, coating peeling, and over-sintering. These defects manifest as abnormal luminescence intensity regions during pyrolysis. The defect evolution model is constructed based on the local incremental change rate as follows:

[0109]

[0110] in,

[0111] : The range of the rate of change of pyrolysis intensity in the i-th and j-th subregions;

[0112] : Data from the output of step S2;

[0113] The range operation indicates the maximum degree of oscillation in the region during the pyrolysis process.

[0114] When a region meets the following conditions:

[0115]

[0116] That is, the difference in the rate of change exceeds the system's set threshold. (The threshold is determined based on the statistical distribution of the difference in the rate of change of each region under historical normal working conditions. Usually, the mean is taken plus multiple standard deviations to ensure that stable regions and potential abnormal regions can be effectively distinguished.) It is determined that the pyrolysis process in this region is unstable and there is a potential risk of defects.

[0117] By using a defect evolution model, "abnormal evolution trends" can be identified in advance before visible cracks or detachment occur, and early warning signals can be generated as triggering conditions for subsequent closed-loop control.

[0118] Traditional anode plate coating monitoring systems often rely on post-processing detection methods, such as plasma scanning, electrochemical testing, or offline image recognition, lacking continuous analysis of the development process and evolution trajectory of coating defects. These methods have significant time lag, making it difficult to identify early abnormal areas in advance, resulting in the inability to intervene in time when the anode plate is over-sintered or under-sintered.

[0119] This step introduces a model based on the extreme difference in pyrolysis intensity evolution rate, which can capture the drastic fluctuations in abnormal regions over time and identify "thermal disturbance" trends in advance. This method is significantly superior to static strategies that only detect intensity anomalies, enabling real-time monitoring and trend warning of defect generation mechanisms, and providing a front-end signal source for closed-loop control of coating process parameters.

[0120] In this embodiment, to improve the robustness and time response capability of the model, this step employs a sliding window algorithm. Dynamic threshold adaptation is implemented, automatically adjusting based on the statistical distribution of the previous batch. This ensures that the criterion maintains consistent sensitivity and specificity across different production cycles.

[0121] To improve the interpretability of the warning, the system will also... Areas exceeding the threshold are dynamically highlighted on the monitoring interface in the form of a heatmap, with an attached trend chart trajectory, allowing operators to intuitively judge the abnormal development path and possible evolution results. This human-machine collaborative anomaly intervention mechanism balances the controllability of automatic identification and human decision-making, significantly enhancing the system's practical value.

[0122] S4. Based on the results output by the defect evolution model, clustering is performed through an irregular clustering mechanism to form an anomaly region map.

[0123] During the pyrolysis of the anode plate, the abnormal area is affected by the uneven micro-coating and the thermal conduction effect of the substrate, resulting in a "band-like" or "sheet-like" structure.

[0124] By using irregular clustering mechanisms to locate abnormal distribution structures, the abnormal distribution structures in step S3 are... Mapped to a global anomaly region map.

[0125] The irregular clustering mechanism is designed with the following clustering sensitivity mapping function:

[0126]

[0127] in,

[0128] : The comprehensive pyrolysis anomaly value of the k-th cluster anomaly region;

[0129] : The set of indexes for all subregions belonging to cluster k;

[0130] Weighting factors are dynamically adjusted based on location, adjacency, and hotspot density.

[0131] The weighting factor is not obtained through preset or manual experience settings, but through dynamic calculation using multi-dimensional spatial information in the anode plate image. Its sources can be divided into the following three dimensions:

[0132] Location-based distribution factor evaluation:

[0133] The location of each anomalous region has a different impact on the overall pyrolysis quality. For example, the central area of ​​the anode plate is often heated evenly, while the edge areas are more prone to cracking due to uneven heat dissipation. Therefore, the system first assigns weights to anomalous regions based on their relative position to the geometric center of the anode plate, with higher weights for regions closer to the center and lower weights for regions closer to the edge, ensuring that high-risk regions are given priority.

[0134] Clustering weight adjustment based on adjacency:

[0135] If anomalous regions exhibit a clear spatial clustering trend, meaning that multiple anomalous clusters are adjacent to or overlap each other, it usually indicates a potential systemic problem (such as local temperature control failure). Therefore, the system dynamically adjusts the weights by calculating the adjacency between anomalous regions (e.g., contact boundary length, number of shared edges, etc.). Anomalous regions with strong adjacency and high cluster density will have their weight factors significantly increased.

[0136] Dynamic weighted gain based on hotspot density:

[0137] By combining temperature field information during the pyrolysis process, density analysis is performed on the distribution of hot spots (hot spots) in the image. If a certain area continuously exhibits high temperature anomalies over multiple time periods, it is designated as a "high-frequency hot spot region," indicating the possibility of abnormal fluctuations or structural defects during the process. The corresponding weighting factor for such regions will be automatically amplified by the system to reflect their significant impact on the overall risk assessment.

[0138] Ultimately, the results from these three dimensions will be uniformly summarized by the system during the image preprocessing and thermal field dynamic analysis stages to form a comprehensive weighting factor specific to each anomaly cluster, which will then participate in the calculation of the subsequent risk index and trend modeling.

[0139] The model is based on comprehensive pyrolysis anomalies. The final output is an anomaly region map of the clustering mapping. The boundaries and intensities of each cluster region in the map are used to guide pyrolysis intervention, including:

[0140] First, for all sub-regions Perform density estimation to identify locally anomalously dense areas;

[0141] Unsupervised density-guided clustering algorithms (such as those based on Local Outlier Factor (LOF) or density peaks) are used to cluster outlier regions, resulting in multiple cluster sets. ;

[0142] The overall cluster strength is calculated by weighting and summing all regions within each cluster according to their anomaly intensity and location weight. ;

[0143] The output is an anomalous region map, and the boundaries and intensities of each cluster are used to guide pyrolysis intervention.

[0144] Existing anode plate pyrolysis monitoring systems generally rely on threshold-based pixel-level detection methods for spatial anomaly identification. Their main drawback is that they cannot identify the "spatial correlation" and "dynamic clustering effect" between abnormal areas. For example, if multiple small thermal disturbance areas are spatially continuous, they are very likely to be different stages of the same defect development. However, existing methods often misjudge them as discrete phenomena and cannot construct an overall defect area map.

[0145] This step innovatively introduces an irregular spatial clustering mapping mechanism to automatically construct the boundaries of anomaly clusters, comprehensively expressing the "regional expansion trend" of anomalies. This clustering strategy not only improves the accuracy and robustness of anomaly detection but also enables subsequent intervention and control to have spatial orientation capabilities, significantly outperforming traditional hotspot analysis methods based on single points or sliding windows.

[0146] S5. Based on the results of the irregular clustering mechanism, a pyrolysis defect intervention and control model is established to guide real-time process adjustment.

[0147] Currently, most mainstream pyrolysis production monitoring systems for chlorine-evolving anode plates remain at the "abnormal alarm" level, lacking an automated linkage mechanism for real-time process intervention in abnormal areas. Even if some adjustment logic exists, it is often based on a linear response of thresholds, such as "cooling down when the abnormal temperature exceeds the limit," without considering multi-dimensional dynamic factors such as the abnormal evolution trend, spatial expansion, and the initial occurrence time of the abnormality.

[0148] Based on existing technology, this step uses the pyrolysis intervention control model output in step S4. The model is shown below as an input parameter:

[0149]

[0150] in,

[0151] : Intervention urgency score for the k-th cluster region;

[0152] Abnormal evolution speed;

[0153] : The pyrolysis initiation delay parameter of the region;

[0154] The weight parameters can be obtained by training with historical samples. The values ​​of the weight parameters are based on historical anode plate defect sample data. They are optimized by cross-validation in the training set by minimizing the deviation between the predicted score and the actual intervention record, so as to ensure that each indicator has a reasonable and representative proportion in the intervention score.

[0155] The pyrolysis intervention control model also includes a limit value for comparison. ;

[0156] when When this occurs, the corresponding control module is automatically triggered, issuing the following intervention strategy command:

[0157] Lower the local pyrolysis temperature;

[0158] Delaying or halting subsequent sintering in a certain area;

[0159] Increase the airflow velocity in this area;

[0160] Adjust the position of the carrier plate to avoid localized overheating.

[0161] This model enables closed-loop control, allowing for efficient coupling of image anomalies with process control, providing not only monitoring capabilities but also prevention and control capabilities.

[0162] S6. Construct a time-period stability scoring model to evaluate the trend of the anode plate production process under real-time monitoring.

[0163] The time-phase stability scoring model is used to analyze the defect control effect of the anode plate in multiple pyrolysis stages:

[0164]

[0165] in,

[0166] : The average defect suppression effect index during the s-th time period;

[0167] : Abnormal values ​​of comprehensive pyrolysis before intervention;

[0168] The overall pyrolysis anomaly value obtained from the reassessment after intervention;

[0169] Avoid using tiny constants with a denominator of zero;

[0170] The total number of abnormal clusters monitored during this period.

[0171] This scoring indicator has three important functions:

[0172] It can quantitatively evaluate whether the intervention measures are effective; make horizontal comparisons of the pyrolysis stability levels of different batches of anode plates; and provide a basis for decision-making on subsequent furnace temperature control strategy adjustments.

[0173] Ultimately, the system outputs "process stability radar charts" and "defect evolution risk distribution charts" divided by batch, realizing a complete closed-loop link from image recognition → defect monitoring → process control → risk modeling.

[0174] Existing methods for assessing the pyrolysis quality of anode plates are mostly based on single-point static image analysis or single-indicator statistics (such as the number of defects or area), lacking longitudinal trend modeling of the entire pyrolysis process and the construction of stage-by-stage indicators. Especially in multi-stage pyrolysis control systems, existing technologies cannot effectively correlate intervention behaviors with quality results, leading to delayed or distorted assessment results.

[0175] The multi-time period and multi-scale time period stability scoring model proposed in this step not only considers the quantitative differences in anomalies before and after pyrolysis, but also introduces the persistence and phased fluctuation characteristics of the intervention effect in the time dimension.

[0176] In this embodiment, different batches of anode plates are often affected by fluctuations in raw materials, differences in electrolytic cell operating conditions, and changes in equipment thermal inertia during actual production, resulting in "semi-repetitive" and "semi-variant" characteristics in the defect morphology identified by the image. If identification is performed from scratch every time, it would waste computing resources and lose the value of reusing existing experience.

[0177] The method also compresses the features of the abnormal region map formed in the pyrolysis images of each batch of anode plate coatings and constructs a defect map fingerprint with high-dimensional vector description;

[0178] The graph fingerprints are persistently stored and assigned batch tags, forming a defect memory bank that is "traceable, transferable, and matchable".

[0179] When subsequent batches of anode plates enter the monitoring process, the system does not perform a complete identification initially, but instead executes a "graph similarity matching operation." This quickly compares the current abnormal clustering graph with existing graph fingerprints. If the similarity is higher than a preset threshold, then:

[0180] Directly invoke the optimal intervention strategy corresponding to this type of historical map;

[0181] There is no need to retrain the model or manually set control parameters;

[0182] Significantly reduces the cost of initial batch debugging and improves system response efficiency.

[0183] This mechanism differs from traditional image classification algorithms or anomaly detection processes. It is the first time that a pattern recognition concept similar to "graphic fingerprinting + experience transfer" has been introduced into the industrial real-time monitoring scenario of chlorine-evolving anode plates. This approach has the following innovative features:

[0184] For the first time, an industrial defect atlas knowledge base has been built, breaking through the inertia of existing image recognition methods that require re-analysis every time;

[0185] Supports real-time matching of historical cases to form dynamic knowledge enhancement paths;

[0186] To avoid repetitive work and realize a new intelligent adjustment mechanism that combines "experience transfer" and "batch linkage".

[0187] Example 2:

[0188] This embodiment also provides a computer device applicable to a real-time monitoring method for the production of chlorine-evolving anode plates based on image processing, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the real-time monitoring method for the production of chlorine-evolving anode plates based on image processing as proposed in the above embodiment.

[0189] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a real-time monitoring method for the production of chlorine-evolving anode plates based on image processing, as proposed in the above embodiments.

[0190] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0191] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0192] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0193] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0194] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0195] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A real-time monitoring method for the production of chlorine-evolving anode plates based on image processing, characterized in that, Includes the following steps: S1. Construct a pyrolysis window image acquisition system. Pre-calibrate the acquired images through thermal field modeling. To establish a quantifiable mapping relationship between luminescence intensity and the real temperature field, the following thermal field calibration model is designed through thermal field modeling: ； in, : The intensity of the fitted thermal field corresponding to pixel (x, y); : The grayscale value of this pixel in the infrared channel; Background reference grayscale base; All of these are calibration constants obtained by pre-fitting the system through a standard blackbody temperature field experiment; The thermal field calibration model is used to ensure that the intensity values ​​of each subsequent pixel have thermophysical meaning in spatial coordinates, thereby constructing the change and evolution structure of the thermal decomposition spectrum; S2. Construct a coating pyrolysis luminescence spectrum model and spatially reconstruct the image calibrated in step S1; in step S2, an image with thermophysical significance has been obtained through the thermal field calibration model established in step S1. Based on the image, a spectrum structure is constructed, and the output from step S1 is... Transformed into time series It also enables spatial scene reconstruction, providing input data structures for subsequent defect identification; The coating surface is divided into multiple grid sub-regions, and a coating pyrolysis luminescence spectrum model is established for the time evolution of each region: ； in, : The average luminous intensity of the i-th and j-th sub-regions at time t; : The pixel size of each grid sub-block; : This refers to the data output in step S1 after the transformation; The coating pyrolysis luminescence spectrum model structures the original image into a regular spatial grid, constructing regional units that are easy to process mathematically; and captures the luminescence change trajectory of the coating at the same location during the pyrolysis process, providing an evolutionary basis for identifying defect points; S3. Based on the data output in step S2, construct a defect evolution model and determine whether there are potential defect risks; defects manifest as abnormal luminescence intensity regions during pyrolysis. The defect evolution model is constructed based on the local incremental change rate as follows: ； in, : The range of the rate of change of pyrolysis intensity in the i-th and j-th subregions; : Data from the output of step S2; When a region meets the following conditions: ； That is, the difference in the rate of change exceeds the system's set threshold. The pyrolysis process in this area was determined to be unstable, posing a potential risk of defects. By using the defect evolution model, the "abnormal evolution trend" can be determined in advance before visible cracks or detachment occur, and an early warning signal can be generated as the trigger condition for subsequent closed-loop control. S4. Based on the results output by the defect evolution model, clustering is performed using an irregular clustering mechanism to form an anomaly region map; S5. Based on the results of the irregular clustering mechanism, establish a pyrolysis defect intervention and control model to guide real-time process adjustment; in step S5, the pyrolysis intervention and control model uses the output of step S4. The model is shown below as an input parameter: ； in, : Intervention urgency score for the k-th cluster region; Abnormal evolution speed; : The pyrolysis initiation delay parameter of the region; Weight parameters; S6. Construct a time-period stability scoring model to evaluate the trend of the anode plate production process under real-time monitoring.

2. The real-time monitoring method for the production of chlorine-evolving anode plates based on image processing according to claim 1, characterized in that, In step S1, the pyrolysis window image acquisition system includes an infrared quartz observation window and an industrial-grade high-resolution infrared camera arranged outside it, used to acquire the light emission image of the surface of the coated anode plate titanium substrate during the pyrolysis process.

3. The real-time monitoring method for the production of chlorine-evolving anode plates based on image processing according to claim 1, characterized in that, In step S3, during the production of the chlorine-evolving anode plate, defects include localized insufficient sintering, coating peeling, and over-sintering.

4. The real-time monitoring method for the production of chlorine-evolving anode plates based on image processing according to claim 1, characterized in that, In step S4, during the pyrolysis of the anode plate, the abnormal area is affected by the uneven micro-coating and the thermal conductivity effect of the substrate, resulting in a "band-like" or "sheet-like" structure. By using irregular clustering mechanisms to locate abnormal distribution structures, the abnormal distribution structures in step S3 are... Mapped to a global anomaly region map.

5. The real-time monitoring method for the production of chlorine-evolving anode plates based on image processing according to claim 4, characterized in that, The irregular clustering mechanism is designed with the following clustering sensitivity mapping function: ； in, : The comprehensive pyrolysis anomaly value of the k-th cluster anomaly region; : The set of indexes for all subregions belonging to cluster k; Weighting factors are dynamically adjusted based on location, adjacency, and hotspot density. Based on comprehensive pyrolysis anomalies The final output is an anomaly region map of the clustering mapping, and the boundaries and intensity of each cluster region in the map are used to guide pyrolysis intervention.

6. The real-time monitoring method for the production of chlorine-evolving anode plates based on image processing according to claim 5, characterized in that, The pyrolysis intervention control model also includes a limit value for comparison. ; when When this occurs, the corresponding control module is automatically triggered, issuing the following intervention strategy command: Lower the local pyrolysis temperature; Delaying or halting subsequent sintering in a certain area; Increase the airflow velocity in this area; Adjust the position of the carrier plate to avoid localized overheating.

7. The real-time monitoring method for the production of chlorine-evolving anode plates based on image processing according to claim 1, characterized in that, In step S6, the time-period stability scoring model is used to analyze the defect control effect of the anode plate in multiple pyrolysis stages: ； in, : The average defect suppression effect index during the s-th time period; : Abnormal values ​​of comprehensive pyrolysis before intervention; The overall pyrolysis anomaly value obtained from the reassessment after intervention; Avoid using tiny constants with a denominator of zero; The total number of abnormal clusters monitored during this period.

8. The real-time monitoring method for the production of chlorine-evolving anode plates based on image processing according to claim 1, characterized in that, The method also compresses the features of the abnormal region map formed in the pyrolysis images of each batch of anode plate coatings and constructs a defect map fingerprint with high-dimensional vector description; The graph fingerprints are persistently stored and assigned batch tags, forming a defect memory bank that is "traceable, transferable, and matchable".

Images