Intelligent printing microcapsule image control system and method

Abstract

The invention relates to the technical field of industrial control systems, in particular to an intelligent printing microcapsule image control system and method.The system comprises an input module for receiving input data, intelligently calculating control parameters and generating an executive command and a control signal, and an execution module for receiving the executive command and the input data and combining real-time state information; the intelligent control module is used for receiving a control parameter and an execution command, outputting a printing stock and related execution data, receiving the printing stock, the control parameter and the execution command, monitoring the color development characteristic in real time, dynamically optimizing and adjusting the control parameter of a divisible region through a self-adaptive control algorithm, and finally outputting the printing stock subjected to color development and detailed execution data; and the optimization module collects the full-process execution data of each module and the quality index for evaluation to obtain an optimization command, and deploys and updates the modules in the system according to the optimization command. The invention discloses a corresponding method, and self-adaptive accurate control is realized through multi-module intelligent linkage.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial control systems, and particularly to an intelligent control system and method for microcapsule images in printing. Background Art

[0002] In industrial manufacturing, advanced imaging materials, such as microcapsules for printing, require subsequent precise programmed excitation to exhibit their final colors and preset functions. The formation of the key characteristics of these materials highly depends on the precise program control of this excitation process.

[0003] The automation of such production processes is usually managed by industrial control systems. When existing industrial control systems execute excitation tasks, they mostly rely on a pre-set fixed program and parameter set, and generally lack the dynamic perception ability and adaptive adjustment mechanism for the real-time state of materials and the dynamic changes in the processing environment. When facing the inherent differences between batches of materials, the slight fluctuations in environmental conditions, and the refined and spatially differentiated requirements for the excitation degree posed by complex images, such fixed-program-based control systems are difficult to achieve high-precision and highly flexible real-time optimization program control. Therefore, in the industrial production using such advanced materials, ensuring their final imaging and functional characteristics and achieving precise control have become a key technical bottleneck currently faced.

[0004] Therefore, an intelligent control system and method for microcapsule images in printing are proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent control system and method for microcapsule images in printing, which realizes adaptive precise control through multi-module intelligent linkage.

[0006] To achieve the above purpose, the present invention provides the following technical solutions: An intelligent control system for microcapsule images in printing, comprising: Input module: Receives the input digital image file as the basic program setting, integrates the inherent characteristic parameters of the selected material and historical operation data, intelligently calculates the relevant control parameters, and generates an execution command and a control signal; Execution module: Receives the execution command and the image setting data, combines the real-time state information monitored by the online sensor, dynamically adjusts the key process operation parameters, and outputs the printing substrate for controlling the deposition of microcapsules according to the instruction and the relevant execution data; Intelligent control module: Receives the printing substrate after deposition processing, the target characteristic parameters and the execution command, real-time monitors the color development characteristics output according to the program through the online monitoring system, judges the deviation between the real-time monitored color development characteristics and the target requirements defined by the target characteristic parameters through the adaptive control algorithm, dynamically optimizes and adjusts the relevant control parameters of the separable regions, and finally outputs the printing substrate that meets the set requirements and the detailed execution data; Optimization module: Collect the execution data and performance indicators of the whole process of each module for evaluation, obtain optimization commands, and deploy and update the modules in the system according to the optimization commands.

[0007] Preferably, for the control parameters related to intelligent computing, the control parameters are control parameters, and the specific acquisition process is as follows: Process the basic program settings, the solid color domain ability data of the selected materials for fusion, and the historical production data through a gradient boosting decision tree model to generate preliminary control parameters; use the preliminary control parameters as the initial solution, and judge according to the color rendering quality optimization goal and process feasibility constraints. If no further optimization is required, the preliminary control parameters are used as the final control parameters after verification and correction; if further optimization is required, use the adaptive perturbation exploration optimization algorithm for iterative refinement and then output the final control parameters.

[0008] Preferably, the adaptive perturbation exploration optimization algorithm includes intelligently selecting the parameters to be adjusted and the perturbation direction based on a heuristic rule set; dynamically and adaptively setting the perturbation amplitude of the control parameters, iteratively generating candidate parameter combinations, and updating the optimal parameter set after verification and evaluation until the termination condition is met.

[0009] Preferably, the dynamic adjustment of key process operation parameters includes: continuously obtaining and analyzing feedback data from on-line sensors that reflect the real-time application state of the current microcapsule material; immediately comparing the obtained real-time application state feedback data with the expected deposition target state received from the control instruction to determine the deviation between the actual application state and the expected deposition target state; According to the deviation, calculate in real time the adjustment amount required for the key deposition parameters, automatically apply this adjustment amount to the corresponding key deposition parameters, and continuously reduce the deviation to make the actual application state of the microcapsule material dynamically approach and maintain at the expected deposition target state.

[0010] Preferably, the real-time calculation of the adjustment amount required for the key deposition parameters includes: Analyze the historical adjustment instruction data of the key deposition parameters and the corresponding real-time material application state feedback data provided by on-line sensors, and use a recursive estimation algorithm to continuously estimate and update the key dynamic characteristic parameters of the microcapsule deposition process on-line; Based on the key dynamic characteristic parameters, automatically and periodically calculate and set the proportional, integral, and differential gain parameters of the PID control algorithm according to the preset PID parameter tuning rules.

[0011] Preferably, the color rendering characteristics output by the on-line monitoring system are monitored in real time according to the program. The color rendering characteristics are specifically color rendering characteristics, including: During the color development process, the image data of each separable area on the printing substrate during the color development process is captured in real time and processed to locate the target area and calculate the quantitative color parameters and color uniformity therefrom, which are used as the color development characteristics of the real-time monitoring output and fed back to the adaptive control algorithm.

[0012] Preferably, the relevant control parameters of the separable area are dynamically optimized and adjusted through the adaptive control algorithm. The relevant control parameters are excitation parameters. The specific processing flow is as follows: for each separable area, the adaptive control algorithm calculates and generates a set of corrected excitation unit parameters aiming to minimize the deviation quantization result and make the color development characteristics approach the target based on the deviation quantization result between the current color development characteristics of the separable area monitored by machine vision in real time and the preset target color development characteristics, in combination with the dynamic process response model. Subsequently, the corrected excitation unit parameters are applied to the excitation unit of the corresponding separable area to adjust its working state in real time.

[0013] Preferably, the execution data and quality indicators of the entire process of each module are collected and evaluated, including: The execution data and quality indicators are processed using the deployment decision model to output the risk assessment indicators and deployment gains. The risk assessment indicators and deployment gains are processed using the decision model in combination with historical deployment cases to obtain the optimization command.

[0014] An intelligent control method for microcapsule images in printing, including: Receiving the input digital image file as the basic program setting, integrating the inherent characteristic parameters of the selected material and the historical operation data, intelligently calculating the relevant control parameters, and generating the execution command and control signal; receiving the execution command and the image setting data, combining with the real-time status information monitored by the online sensor, dynamically adjusting the key process operation parameters, and outputting the printing substrate on which the microcapsule deposition is regulated according to the instruction and the relevant execution data; receiving the printing substrate after the deposition treatment, the target characteristic parameters and the execution command, monitoring the color development characteristics output in real time according to the program through the online monitoring system, judging the deviation between the color development characteristics monitored in real time and the target requirements defined by the target characteristic parameters through the adaptive control algorithm, dynamically optimizing and adjusting the relevant control parameters of the separable area, and finally outputting the printing substrate that meets the set requirements and the detailed execution data; collecting the execution data and performance indicators of the entire process of each module for evaluation to obtain the optimization command, and performing deployment updates on the modules in the system according to the optimization command.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. By integrating digital images, the color gamut capabilities of materials, and historical production data, and using intelligent algorithms to calculate and optimize the initial color rendering parameters, a highly optimized and task-specific initial setpoint is provided for the entire precise control process from the source. This ensures that the subsequent physical execution and color rendering process are carried out under the guidance of ideal state parameters, greatly reducing the initial error and laying a high-quality preset foundation for achieving the final precise imaging and functional characteristics, which is the key to realizing forward-looking precise control.

[0016] 2. By using an online sensor to continuously monitor the actual state during the microcapsule deposition process and combining intelligent control algorithms such as parameter self-tuning PID to dynamically adjust the key process operation parameters, real-time, closed-loop precise control of the microcapsule physical deposition process is achieved, ensuring that the actual deposition result of the microcapsules matches the requirements in the execution command.

[0017] 3. By using machine vision to perform real-time, quantitative online monitoring of the color rendering characteristics of the printed substrate after deposition and dynamically adjusting the excitation parameters of separable regions through an adaptive control algorithm, refined, closed-loop adaptive precise control of the key functional formation stage of microcapsule color rendering activation is achieved. This ensures that even in the presence of cumulative deviations in the previous process and uncertainties in material responses, the final color rendering effect can be precisely guided to the target characteristics.

[0018] 4. By collecting and deeply analyzing and evaluating the execution data and final quality indicators of the entire system's full process, using an online learning mechanism to generate optimization commands to update the control models and strategy parameters of each module, a system-level continuous learning and self-optimization ability is established, thereby ensuring and continuously improving the long-term precise control performance of the entire system and its adaptability to changes. This overcomes the problem of the decline in precise control ability caused by model aging and non-optimal parameters in traditional systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 FIG. is a schematic structural diagram of an intelligent control system for microcapsule images used in printing provided by an embodiment of the present invention; Figure 2 FIG. is a schematic flow diagram of an input module provided by an embodiment of the present invention; Figure 3 FIG. is a schematic flow diagram of a method for intelligent control of microcapsule images used in printing provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0021] Please refer to Figures 1 to 3 , the present invention provides an intelligent control system and method for microcapsule images for printing, and the technical solutions are as follows: Embodiment 1: In order to achieve intelligent and precise control of microcapsule images, Company A introduced an intelligent control system for microcapsule images for printing provided by the present invention. The system structure diagram is as Figure 1 shown, and the specific process is as follows: Input module: Receive the input digital image file as the basic program setting, integrate the inherent characteristic parameters of the selected material and the historical operation data, intelligently calculate the relevant control parameters, and generate execution commands and control signals. The flow schematic diagram of the input module is as Figure 2 shown.

[0022] The operator inputs the digital image file to be printed into the system through the interaction interface. The input module receives the input digital image file and parses its content (pixel data, size, color profile, etc.) into the basic program setting that defines the content and basic requirements of the subsequent printing operation. The basic program setting encapsulates all the necessary information for precisely defining and guiding the core process of subsequent microcapsule image printing. In this embodiment, the basic program setting specifically consists of pixel-level color data, spatial dimension and resolution parameters, color space definition information, and additional structure and metadata information. The additional structure and metadata information include Alpha channel data for defining the transparency of each part of the image, layer structure information existing in specific file formats (such as layered TIFF or PSD files), and descriptive metadata such as image titles or identifiers, creation dates, author information, and copyright statements that are helpful for task identification, content confirmation, and production traceability.

[0023] Further, according to the printing microcapsule material and substrate type selected by the operator through the interaction interface when starting the printing task, the input module will retrieve the corresponding solid color gamut ability data from the built-in and calibrated material database in the system. Then, historical production data is obtained, and missing items, abnormal data points caused by various reasons, and common noise interferences in sensor data that may exist in the data records of the historical production data are identified and processed. Then, different numerical features are standardized, and categorical information is effectively numerically encoded; features with extremely small changes are removed through variance analysis. On this basis, the input module continues to execute the data fusion step. This step uses the technical means of feature vector splicing to integrate the key extracted features of the current digital image, the quantified solid color gamut characteristics data of the selected material, and the preprocessed effective historical production experience data that have been separately obtained and processed before into a unified, multi-dimensional high-dimensional feature vector according to a predetermined structure. The constructed high-dimensional feature vector will be used as the direct and standardized input for the subsequent gradient boosting decision tree algorithm.

[0024] Further, after the data fusion step is completed, the input module immediately submits this high-dimensional feature vector integrating the current image analysis features, the solid color gamut ability data of the selected material, and the preprocessed historical production data information as a direct input to a pre-trained gradient boosting decision tree (GBDT) model. After receiving the high-dimensional feature vector, the GBDT model performs a prediction processing operation. The core of this prediction processing operation is that the model uses the complex decision logic learned and solidified from a large amount of historical production data during the training stage: in the numerous decision trees integrated inside the model for the input high-dimensional feature vector, layer-by-layer path selection and conduction are performed according to the feature judgment rules learned on each tree node. The analysis results of each tree are integrated through an integration mechanism, so as to jointly infer the output parameters that best match the current input features. Finally, the GBDT model generates a set of structured preliminary control parameters, and the control parameters are specifically color development parameters. This parameter set stipulates the initial microcapsule material deposition quantification setting and the preliminary color development process excitation condition setting for each preset analysis area of the current input digital image, serving as the starting input for the subsequent optimization and refinement process.

[0025] After obtaining the preliminary control parameters, the input module starts the optimization and refinement process. Taking these preliminary parameters as the initial solution, and strictly in accordance with the color rendering quality optimization objectives set for the current specific printing task and various process feasibility constraints, the process first conducts a strict process feasibility verification on this preliminary parameter set. If any parameter value is detected to violate the preset key hard constraint conditions, such as the upper and lower limits of the material usage or the safety range of the color rendering excitation energy, etc., the algorithm will first initiate an efficient constraint correction program. This program uses a targeted parameter adjustment strategy with minimal impact, which can directly correct the out-of-bounds parameter value to the nearest allowable boundary, thus quickly correcting the entire parameter set to be completely within the process feasible region, ensuring that the basis for subsequent optimization iterations is effective and executable.

[0026] Furthermore, for the preliminary parameter set that has been ensured to meet the hard constraints, the system then conducts a lightweight performance estimation and pattern comparison. The system calls a simplified color difference prediction function module constructed based on the core theory of color science. The core of this module is a preset, non-iterative forward color space conversion model; specifically, the deposition amounts of each microcapsule channel and the corresponding preliminary excitation energy values in the current parameter set are mapped to the CIELAB color space. To achieve high-speed calculation, this conversion model adopts the structure of a small three-dimensional lookup table. Subsequently, the module uses a fast evaluation that optimizes the operation of the standard color difference calculation formula to compare the predicted color values output by the aforementioned forward color space conversion model with the target color values represented by the average color features of several key regions extracted from the current input digital image file. Through this series of precise and efficient calculation processes, this function module finally quickly calculates and outputs the expected color difference quantization index for the said key image regions. The overall algorithm logic of this prediction function module and all the conversion parameters it requires are set in the system as fixed configurations.

[0027] Furthermore, the system compares the values of the key parameters of the current preliminary parameter set with a pre-established and maintained historical robust parameter pattern library. This pattern library is constructed through in-depth offline statistical analysis of a large amount of accumulated historical successful production case data. This pattern library specifically and clearly defines the effective value ranges, their statistical central tendencies, and common parameter combination patterns of the control parameters that can continuously and stably produce high-quality printing outputs under preset conditions such as material types, substrate characteristics, and classification features extracted from image content under similar printing conditions. The comparison process is achieved by calculating the statistical distance metric between the current preliminary parameter set and these historical robust patterns. This metric can specifically be to determine whether the current parameter values fall within the high-frequency confidence intervals corresponding to the historical data, so as to evaluate the normality of the current parameter set and the potential process stability risks.

[0028] The system will execute the final optimization trigger decision based on the expected color difference quantization metrics, compliance, and distance metric results of each key image region in the comprehensive output. If the expected color difference quantization metrics of all key image regions have met the core color accuracy index requirements set by the task, and the overall pattern of the current preliminary parameter set shows a high degree of compliance with the corresponding pattern in the historical robust parameter pattern library, the system comprehensively determines that the preliminary parameter set does not require further iterative optimization. In this specific case, the preliminary parameter set will be directly adopted as the final control parameter and prepared for output to the subsequent relevant processing modules. Conversely, if the expected color difference index fails to meet the standards comprehensively, or the current parameter pattern significantly deviates from the robust region indicated by historical experience, or both adverse situations exist simultaneously, the system determines that the current initial control parameter needs to be further optimized to improve the final printing quality and reduce potential production risks. It will trigger and start the adaptive perturbation exploration optimization algorithm for optimization.

[0029] Through the built-in and preferentially executed hard constraint verification and immediate correction function, ensure that all preliminary parameters first meet the basic process feasibility. Then, through the comparison and analysis of a lightweight efficiency estimation module with the historical robust parameter pattern, quickly and intelligently judge the parameter quality. This enables the system to identify and directly adopt parameters that are already good enough, avoiding unnecessary in-depth iterative optimization of all parameters, thus significantly shortening the average processing time and quickly providing high-quality initial working set points for the subsequent modules.

[0030] Furthermore, starting the adaptive perturbation exploration optimization algorithm for optimization includes, through a preset heuristic rule set of microcapsule printing process rules, combining the deviation characteristics between the color rendering effect estimated by the current parameter set and the color rendering quality optimization target, intelligently selecting a set of control parameters that have the most significant current deviation on the overall printing quality from the current parameter set as the adjustment object for this round of iteration. At the same time, the heuristic rule set provides a perturbation direction with a clear improvement intention for the adjustment of these selected parameters. When the algorithm evaluates and finds that there is a specific deviation between the expected color rendering effect of the current preliminary parameters and the target, the algorithm calls the rule set to guide the adjustment of the parameters.

[0031] Table 1 Example fragment of the APEO algorithm heuristic rule set

[0032] Table 1 shows an example of a fragment of the APEO algorithm heuristic rule set. These rules are based on process laws and color science principles and are used to guide the algorithm on how to intelligently select adjustment parameters and adjustment directions based on current deviations when optimizing control parameters. For example, if the estimated color is significantly lower than the target value in brightness, the rules will guide the algorithm to prioritize and try to increase those control parameters that have the most direct positive impact on the overall brightness (such as a key excitation energy or the amount of a specific microcapsule). Similarly, for specific hue or saturation deviations, the rule set will also guide the algorithm to adjust the color channel parameters that are most relevant to it and clarify the adjustment direction (increase or decrease) in order to most effectively correct the deviation. At the same time, the rules will also combine the distance between the current value of the parameter and the process constraint boundary to affect the priority and direction of the adjustment to ensure the safety of the adjustment.

[0033] Furthermore, for the selected parameters and their determined perturbation directions, the algorithm uses a dynamic adaptive mechanism to set the specific value of the parameter that should be adjusted in this iteration, that is, the perturbation amplitude. This dynamic adaptive mechanism will first evaluate the current stage of overall optimization: if it is judged that it is in the early stage of optimization or the current solution is still significantly different from the preset quality target, it tends to set a relatively large baseline perturbation amplitude to encourage broader exploration in the parameter space. At the same time, the mechanism will closely track and analyze the actual feedback of recent iterations: if continuous perturbations can bring more significant quality improvements, the perturbation amplitude may be maintained or moderately increased to accelerate optimization; on the contrary, if the improvement effect is not obvious or stagnates, it will clearly indicate to significantly reduce the perturbation amplitude, guiding the algorithm to enter a more sophisticated local search and fine-tuning mode. In addition, for each parameter selected for perturbation, the mechanism will strictly consider the distance between its current value and the upper and lower limits defined by its process feasibility constraints. When the parameter value is very close to any constraint boundary, the rule will forcibly reduce any perturbation amplitude that may point to or exceed the boundary, and even set it to zero in extreme cases to ensure that all perturbation attempts are prioritized to remain within the feasible domain or its immediate edge. Through this set of adaptive rules that integrate the optimization stage, recent results, and constraint boundary considerations, the algorithm determines the precise and dynamically changing perturbation amplitude for each selected parameter in the given perturbation direction.

[0034] According to the specific disturbance amplitude adaptively determined for each selected parameter and its disturbance direction, the algorithm applies these calculated adjustments with clear directions and adaptive amplitudes to the corresponding selected parameters based on the currently recorded optimal parameter set, while other unselected parameters remain unchanged. Through this operation, the algorithm generates new candidate parameter combinations. For the newly generated candidate parameter combinations, all preset process feasibility constraints must be checked immediately. Any candidate solution that fails to fully meet all hard constraints will be judged as infeasible and directly discarded without entering the subsequent quality assessment step.

[0035] Furthermore, for all valid candidate solutions that pass the constraint verification, the system will, based on the preset chromaticity quality optimization objective function, which mainly calculates the standard color difference (ΔE value) between the predicted color and the target color, comprehensively consider other key visual characteristics such as color uniformity and saturation, and output a comprehensive performance evaluation value to guide the iterative selection of the optimization algorithm. The algorithm compares the performance evaluation values of all valid candidate solutions with the evaluation value of the currently recorded global optimal solution. If there is a valid candidate solution with better performance, the algorithm will adopt the candidate solution that results in the most significant improvement in the quality objective function value to update and replace the currently recorded global optimal parameter set. If all tentative perturbations in this iteration fail to produce a valid new solution better than the current optimal solution, the current optimal parameter set remains unchanged in this iteration. This iterative process including the above steps will continue until a clear algorithm termination condition is met. These termination conditions are specifically specified as follows: the algorithm has reached the maximum allowable number of iterations set for the current printing task; in several consecutive iterations, the quality objective function value of the optimal solution has not produced a meaningful improvement greater than the preset threshold, indicating that the optimization process may have converged; the cumulative execution time of the entire APEO optimization process has reached the specific time budget allocated for this step.

[0036] When the APEO algorithm ends its optimization process due to any termination condition, its final output is a set of final control parameters that have been fully refined through this adaptive perturbation exploration optimization process and that maximize the preset chromaticity quality objective under the premise of meeting all process constraints.

[0037] Based on the determined final control parameters, combined with the image geometric information, necessary region division data, and relevant image content features obtained from the basic program settings of the printing task, these multi-source information are compiled, calculated, and integrated. Through this process, the module constructs a structured, machine-readable execution command. To ensure the coordinated operation of all parts of the system, the input module is also responsible for generating and distributing control signals. The main function of these signals is to activate the subsequent execution module and intelligent control module according to the preset logic, notify them that the relevant execution commands and necessary image data are ready and available for invocation, and guide them to start the precise deposition process of microcapsules and subsequent chromaticity control and monitoring tasks according to the established operation process.

[0038] Through heuristic rules that deeply embed the laws of printing processes and the principles of color science, intelligent guidance for the parameter adjustment direction and objects is achieved, combined with a dynamic adaptive mechanism to precisely control the perturbation amplitude. This design enables it to efficiently and targetedly perform optimization searches in complex parameter spaces, thus ensuring that the final control parameters output not only strictly meet all process feasibility constraints but also achieve significant improvements in multiple key visual quality dimensions such as color difference, color uniformity, and saturation. Importantly, the core optimization logic of this algorithm does not rely on additional online machine learning model training but is based on preset intelligent rules and adaptive adjustments. Therefore, APEO can provide a significantly optimized and high-quality initial working set point for the subsequent real-time closed-loop control module within a reasonable computing time, which is crucial for enhancing the response speed, stability, and consistency and quality of the final product of the entire system.

[0039] Execution module: Receives execution commands and image setting data, combines with the real-time status information monitored by online sensors, dynamically adjusts key process operation parameters, and outputs the substrate for controlling the deposition of microcapsules according to the instructions and related execution data.

[0040] The execution module receives the part of the execution command directly related to the microcapsule deposition operation in the previously generated execution command, and at the same time receives the necessary and processed image data. The execution command details information such as the target deposition amount, spatial positioning coordinates, and deposition order of each color microcapsule for each image area in the current printing task. The image data provides a position reference for the visual content during the deposition process. The execution module integrates and controls a piezoelectric nozzle array capable of spraying trace amounts of microcapsule materials as needed, and a specific nozzle group is responsible for a specific color of microcapsules. At the same time, the execution module also controls the precision conveying and positioning subsystem of the substrate to ensure that the relative position between the substrate and the deposition unit is precisely controllable. After receiving the execution program, the execution module first drives the deposition unit to start depositing the calculated target amount of microcapsule material on the specified micro-region of the substrate according to the initial instructions in the program.

[0041] Further, to achieve dynamic closed-loop control of the deposition process, the execution module incorporates an online sensor. The online sensor continuously monitors the real-time application status of the microcapsule material on the substrate. The sensor includes a high-speed, high-resolution vision sensor installed near the deposition unit for capturing in real time the geometric morphology, position accuracy, and coverage density or filling uniformity in a specific micro-region of the just-deposited microcapsule lattice. Then, it directly outputs to the execution module the quantified physical data representing various key deposition states that have been preliminarily processed or calibrated. After receiving these quantified data streams from different sensors, the execution module mainly performs numerical scale unification processing on them. This processing step aims to eliminate the numerical scale differences caused by different physical dimensions of the sensor data and ensure that all state information has a consistent and comparable magnitude before entering the subsequent control algorithm. The state parameters after this unified scale processing are then integrated and dynamically updated into a structured data record, which constitutes the real-time state information and is used to accurately and consistently characterize the actual deposition situation of the current microcapsules for direct invocation by the subsequent closed-loop control algorithm.

[0042] Further, the acquired structured real-time state information (which includes quantified feedback data such as the actual microcapsule application amount, position coordinates, and coverage uniformity in each micro-region) is compared region by region with the desired deposition target state extracted from the corresponding instruction part of the currently executing execution command (i.e., the target application amount, target position coordinates, target uniformity, etc. set for the same micro-region in the program instruction). Through this comparison, the system accurately calculates the deviation vector between the actual application state and the desired deposition target state.

[0043] The deposition target is clearly defined through precise program instructions, and the specific deviation between the actual deposition state and the desired target is accurately quantified through the standardized online sensor feedback. This clear and quantified deviation vector is the key prerequisite and direct driving force for the subsequent intelligent control algorithm to make effective and precise adjustments to ensure the final deposition quality and consistency, and also accumulates important data for the overall learning and optimization of the system.

[0044] The PID controller uses the received deviation vector as its core error input. For each currently controlled key process operation parameter, the controller multiplies its current error, the integral term of the error accumulated over time, and the rate of change of the error over time by their respective proportional (Kp), integral (Ki), and derivative (Kd) gain parameters. Subsequently, the controller linearly superimposes these three separately calculated control action terms to calculate in real time a specific adjustment amount for this key process operation parameter. The adjustment amounts calculated by the PID controller directly act on the key process operation parameters, which are the adjustable characteristics of the precise electrical pulse waveform that drives the microcapsule injection unit and directly affects the physical deposition behavior of the microcapsules, the relevant set values that control the supply flow rate of the microcapsule material, and the fine-tuning instructions for the servo motor system responsible for the relative movement and precise positioning between the substrate and the deposition nozzle.

[0045] Furthermore, to ensure that the PID controller always maintains optimal control performance in the face of possible dynamic changes in process conditions during printing, the system is equipped with a parameter self-tuning function for the PID controller. This parameter self-tuning function mainly includes the following two core sub-processes that work together. The system continuously monitors and records the historical adjustment instruction sequence applied to each key process operation parameter and the precise feedback data sequence provided by the online sensor system that is synchronized with it regarding the real-time application state of the microcapsules. The system uses the recursive least squares method to estimate and update a set of key dynamic characteristic parameters used to describe the microcapsule deposition process based on the latest input-output data. These dynamic characteristic parameters effectively characterize the equivalent gain, main time constant, pure dead time, and discrete model coefficients of the process model under the current operating conditions. Based on the latest set of dynamic characteristic parameters, the proportional (Kp), integral (Ki), and derivative (Kd) three core gain parameters of this PID control algorithm are calculated and set automatically according to the internal model control (IMC) at a preset period and triggered when significant changes in the monitored process dynamic characteristics occur.

[0046] The adjustment amounts for each key process operation parameter calculated in real time by the PID controller will be automatically and precisely applied to the corresponding physical actuator through the control interface. The complete closed-loop control process including deviation determination, adjustment amount calculation, and adjustment amount application will continuously and rapidly control each monitored micro-region deposition unit, enabling the actual application state of the microcapsule material to continuously and dynamically approach and finally precisely and stably maintain the desired deposition target state specified by the execution command under the influence of various potential disturbance factors.

[0047] By continuously online identifying the dynamic characteristics of the deposition process and adaptively adjusting the PID core gain parameters accordingly, the key technical advantages are achieved: it can significantly improve the absolute accuracy and long-term consistency of microcapsule deposition, effectively suppress various unmodeled dynamics and external disturbances during the production process, thus ensuring a high degree of stability and reliability of the deposition quality under complex actual working conditions. At the same time, this mechanism greatly reduces the need for operators to perform frequent and experience-dependent manual PID parameter debugging, significantly improving the automation level of the system, the production preparation efficiency, and the adaptability to diverse printing tasks.

[0048] Intelligent control module: Receives the printed substrate after deposition, control parameters, and execution commands, and through the online monitoring system, it monitors the color development characteristics in real time according to the program. It judges the deviation between the color development characteristics monitored in real time and the target color characteristics defined by the control parameters through the adaptive control algorithm, dynamically optimizes and adjusts the control parameters of the separable regions, and finally outputs the printed substrate with completed color development and detailed execution data.

[0049] The intelligent control module first receives from the execution module the printed substrate with the microcapsule pattern that has been accurately deposited but not fully developed. At the same time, the module retrieves from the execution commands generated by the input module the control parameters related to the current printing task and the initial excitation settings for each separable region. The control parameters clearly define the target color characteristics that the image area is ultimately expected to achieve, for example, the target CIELAB L*a*b* coordinate values, the acceptable color difference ΔE range, and the color uniformity standard. The initial excitation settings provide a starting working point for the subsequent color development process. According to the color development mechanism of the used microcapsules, for example, photosensitive curing color development, thermally activated color development, and laser selective color development, this excitation unit can specifically be an array of high-precision adjustable ultraviolet LED light sources, an infrared heating matrix, and a laser emission system controlled by a precision scanning galvanometer. These excitation units can apply precisely controlled energy, such as light intensity, wavelength distribution, irradiation duration, heating temperature, and duration, to specific separable regions on the printed substrate according to the control instructions.

[0050] Furthermore, while the color development excitation process is started and continuously carried out, the on-line monitoring system is specifically a machine vision system. The machine vision system, in accordance with the monitoring program and area division defined in the execution command, continuously and real-timely monitors the color development status of each separable area on the printing substrate. The machine vision system dynamically adjusts the frequency according to the color development rate, and real-timely captures the image data of the color gradually formed on the surface of the printing substrate under the excitation effect. These image data are then sent to the image processing unit. The image processing unit first performs noise removal and geometric correction processing on the captured images, and then accurately locates the positions of each separable area to be monitored in the current field of view. Then, from these located areas, the average CIELAB L*a*b* value of the current color development status and the standard deviation of the color difference within the area are calculated. These calculated L*a*b* and the standard deviation of the color difference within the area constitute the color development characteristics of real-time monitoring, and as high-frequency feedback signals, they are continuously sent to the subsequent adaptive control algorithm.

[0051] By providing accurate printing targets and initial excitation settings, combined with the use of an advanced machine vision system that can perform real-time and quantitative color and uniformity feedback on controllable partitions and dynamically adjust the monitoring frequency according to the color development rate, it lays a technical foundation for the subsequent adaptive control algorithm to accurately judge the current deviation and then dynamically optimize the excitation parameters to achieve the final accurate color reproduction and high consistency.

[0052] Furthermore, for these specific deviation amounts between the current actual color rendering characteristics (this characteristic covers the average CIELAB value and the color uniformity index) of each separable region and the target color characteristics, which are real-time feedback by the machine vision system, the adaptive control algorithm activates its core dynamic excitation parameter optimization and adjustment process. This process precisely utilizes a dynamic process response model with prediction capabilities that is currently deployed in the system. In its internal structure, this model accurately represents the dynamic influence relationship of the adjustment of various excitation parameters within a specific separable region on the expected evolution of the color rendering characteristics in the short term, especially the quantitative prediction of the evolution trajectory of the L*a*b* color coordinates. In each control cycle of the intelligent control module, when new deviation information is received, the adaptive control algorithm first constructs a set of candidate correction excitation unit parameter adjustment schemes using heuristic logic based on the specific deviations (covering their nature and magnitude) between the color rendering characteristics of each separable region and the corresponding target color characteristics that are currently known. First, this logic analyzes each deviation to identify the dominant deviation component that currently has the most significant impact on the overall color quality (according to the preset quality optimization goal). For example, is the deviation of brightness L* the largest, or is the deviation of a certain chromaticity channel a* or b* the most prominent, or is the color uniformity index the least ideal? After determining the dominant deviation component, the core of this heuristic logic is to refer to and execute an internally defined and pre-constructed control parameter adjustment heuristic rule set. The construction of this rule set does not rely on machine learning training during the online operation of the algorithm, but is comprehensively formulated and solidified in multiple aspects during the system design stage. First, it includes a deep physicochemical understanding of the specific microcapsule color rendering mechanism adopted. For example, how different excitation energies (light intensity, wavelength, heat) affect the reaction rate of microcapsules, the production amount of color-developing substances, and the final color state. Second, it includes the basic principles of recognized color science. For example, in color spaces such as CIELAB, the corresponding relationship between the changes of each component (L*, a*, b*) and color perception (brightness, red-green, yellow-blue), and the basic laws of color mixing. In addition, it also includes systematic experimental data analysis conducted offline and the experience induction of historical successful production cases. From a large amount of experimental and production data, through statistical analysis and expert summary, effective adjustment strategies and parameter sensitivity information for typical deviation patterns are refined. This control parameter adjustment heuristic rule set specifically and logically encapsulates different types of color rendering characteristic deviations. For example, the clear mapping relationship between the identified dominant deviations such as the L* value being significantly lower than the target, the a* value being overly red, or the poor color uniformity of a certain image area and the key excitation parameters recommended for adjustment, and the clear mapping relationship between the total irradiation duration, peak power of a specific light source, the energy ratio of light sources with different wavelengths, and the key points of the temperature curve in a specific heating area. It presets the recommended adjustment directions and adjustment priorities for these parameters aimed at neutralizing the current dominant deviation.According to this rule set, when the algorithm identifies the current dominant deviation, it can automatically query and match the corresponding adjustment rules, thereby identifying the key excitation parameters that need to be adjusted and obtaining a preliminary adjustment direction with a clear improvement intention.

[0053] Furthermore, for each identified key excitation parameter, instead of conducting a broad and aimless search, the algorithm systematically generates several specific and small-scale tentative adjustment values around the current set value of this parameter. These adjustment values include an option to slightly increase the parameter value, an option to slightly decrease the parameter value, and in some cases, may also include an option to keep the parameter value unchanged. Applying these specific tentative adjustments for the selected key excitation parameters separately to the current overall excitation parameter setting forms different candidate modified excitation unit parameter adjustment schemes within the small neighborhood of the current solution. These precisely generated candidate adjustment schemes will then be submitted to the dynamic process response model. Subsequently, taking each candidate modification scheme as input and invoking the dynamic process response model, the expected time evolution trajectories of the color rendering characteristics in the corresponding region within the next few control steps after applying this specific adjustment are independently predicted for each candidate scheme.

[0054] Next, the system uses a comprehensive dynamic trajectory utility function to conduct a quantitative utility evaluation and ranking of these multiple color evolution trajectories predicted by the model. The operating mechanism of this utility function is as follows: First, for each predicted color evolution trajectory, the function precisely calculates and extracts a set of predefined key dynamic performance indicators (KPIs) from it. These indicators specifically describe various important qualities of the trajectory, mainly including: the final predicted color difference, that is, the color difference between the color reached at the end of the predicted trajectory sequence and the target color; the expected convergence time, which measures the number of control cycles required for the predicted color trajectory to reach and stabilize within the error band allowed by the target color; the maximum overshoot of the color component, referring to the maximum deviation amplitude by which any color component (such as L*, a*, b*) exceeds its target value during the approach to its target value; and the smoothness of the trajectory change, which quantifies the smoothness of the color change by evaluating the oscillation degree or irregular fluctuation amount of the predicted trajectory.

[0055] Further, after calculating these independent dynamic performance metrics, the comprehensive dynamic trajectory utility function combines these performance metrics from different dimensions into a single total score representing the comprehensive utility and desirability of the predicted trajectory through weighted linear combination. Among them, the weight coefficients corresponding to each dynamic performance metric reflect the performance preferences and priorities of the currently preset quality mode of the system for various dynamic qualities (such as pursuing the fastest color development speed, the highest color accuracy, and the best process stability). These weight coefficients are applied as the currently effective control strategy parameters to clarify the relative importance of each dynamic quality in the calculation of the comprehensive performance score. Finally, the algorithm selects the parameter of the correction excitation unit corresponding to the expected color evolution trajectory that can generate the highest total comprehensive utility score from all candidate correction schemes that have undergone prediction and utility evaluation as the output instruction generated in this control cycle for real-time adjustment of the working state of the excitation unit.

[0056] By using the dynamic process response model to predict the short-term color evolution trajectories of multiple candidate parameter adjustments and selecting the best through a heuristic trajectory quality evaluation function that considers dynamic qualities such as convergence speed, stability, and smoothness, the adaptive control algorithm realizes the forward-looking intelligent optimization of the excitation parameters. It not only effectively corrects the current deviation but also guides the color development process to quickly and stably approach the target along the optimal path, thus significantly improving the accuracy of color control, dynamic quality, system adaptability, and the efficiency of online decision-making.

[0057] Optimization module: Collects the execution data and quality indicators of the entire process of each module for evaluation, obtains the optimization command, and deploys and updates the modules in the system according to the optimization command.

[0058] The optimization module is configured with a dedicated data interface and processing unit to continuously and real-time collect and aggregate the execution data and result data covering the entire process of each microcapsule image printing task from all other business modules in the system (i.e., the input module, the execution module, and the intelligent control module). These data specifically include: Summary of the key features of the original image file, the preliminary control parameters generated by the GBDT model, the decision results of the conditional optimization gate, the final control parameters output by the APEO algorithm, and the core content of the generated execution commands. During the microcapsule deposition process, the actual set values and adjustment history of each key process operation parameter, the real-time deposition status feedback data sequence monitored by on-line sensors (such as vision sensors), the internal state of the parameter self-tuning PID controller (such as the identified process model parameters, the adjusted PID gains), and any deviation and correction records that occurred. During the color development process, the color evolution trajectories of each separable region (such as L*a*b* value sequence, uniformity index change) monitored in real time by the machine vision system, the dynamic adjustment history and specific values of the excitation parameters by the adaptive control algorithm, and the compliance of the color characteristics of each region with the target when the final color development is completed. The indicators automatically obtained by the system-integrated on-line quality inspection equipment and the comprehensive evaluation results of the final printed matter quality manually entered after off-line inspection, such as the overall verification color difference ΔE based on the standard color card, the visual defect records of specific image regions, and the customer feedback level. All these multi-source heterogeneous data are assigned a unique task identifier, a timestamp, and relevant context information (such as the materials used, the substrates, the equipment status). After cleaning, transformation, and alignment, they are uniformly stored in a structured central historical production database, providing a high-quality data basis for subsequent evaluation and learning.

[0059] The optimization module is triggered by the operator or set at a fixed cycle time to perform in-depth analysis and performance evaluation on the data accumulated in the historical production database, monitor the long-term performance trends and stability of the overall system and each key module (such as GBDT parameter prediction, APEO optimization efficiency, PID deposition control accuracy, intelligent color development control effect), and identify whether the applicability of the current system-level predictive control model (the GBDT model of the input module, the dynamic process response model used by the intelligent control module) is weakened due to changes in process conditions. Analyze the effectiveness of the existing adaptive control strategies (the heuristic rule parameters of the APEO algorithm, the trajectory evaluation function weights of the intelligent control module, the self-tuning rule meta-parameters of the PID in the execution module), and judge whether there is room for further optimization. Through correlation analysis, identify the key influencing factors that cause quality problems and efficiency bottlenecks. Based on the above evaluation results, the optimization module can identify the specific models and parameter sets that most need to be optimized and adjusted in the current system and determine the optimization goals and directions.

[0060] Having determined the optimization requirements and objectives, the optimization module will call its core online learning mechanism and utilize the accumulated and preprocessed historical production data to dynamically optimize and retrain the system-level predictive control model and adaptive control strategy that need to be optimized. For the predictive control model (GBDT, dynamic process response model), the optimization module can adopt an incremental learning algorithm to fine-tune the online parameters of the existing model using the newly obtained data; when the model performance drops to a certain extent, a comprehensive offline retraining based on an extended dataset is triggered to generate an optimized model.

[0061] Regarding the adaptive control strategy adopted within the system and its core adjustable parameters, such as the heuristic rule weights that may exist in the APEO algorithm, the weight coefficients of the trajectory evaluation function in the intelligent control module, and the tuning rule meta-parameters in the parameter self-tuning PID control algorithm of the execution module (such as the expected response time, robustness index weight), the optimization module uses a genetic algorithm for offline optimization. This genetic algorithm iteratively evaluates and screens numerous possible combinations of these strategy parameters. In each generation of evolution, the algorithm uses the system comprehensive performance (e.g., average printing quality, production efficiency, material consumption over a period of time) obtained from the simulation and evaluation based on historical data as the fitness function, retains and combines the parameter sets with better performance, and introduces random mutations to explore new possibilities until a set of strategy parameter settings that can significantly improve the expected overall performance of the system is found.

[0062] After generating the optimized model and strategy, to ensure that their introduction will not have a negative impact on system stability and production quality, the optimization module first conducts internal verification. The effectiveness of the optimized model and strategy is verified through performance evaluation on the historical data backtest set. Only the optimized model and strategy that pass the verification, are confirmed to have better performance than the current version, and are stable and reliable will enter the automatic deployment process. A pre-trained deployment evaluation model is called, which processes the data characterizing the characteristics of the optimized model and strategy (covering its expected performance and resource occupancy) and the data characterizing the current system state and potential deployment impacts to output a set of quantified deployment condition evaluation indicators.

[0063] For the optimization results passed the internal verification, the optimization module then calls a pre-trained multi-factor weighted scoring model specifically for quantitatively evaluating the deployment conditions. The internal structure of this scoring model and the weight coefficients of its various evaluation factors are established by the optimization module through offline machine learning and statistical modeling analysis using a large amount of historical deployment data. These historical deployment data detailed record the characteristic parameters of the newly deployed models and strategies, the specific system state information at the time of deployment, and the actual impact feedback data on the overall system performance and operation stability after deployment. When running, the multi-factor weighted scoring model receives the detailed characteristic data of the new version model and strategy to be deployed currently. This characteristic data clearly covers the expected performance improvement range demonstrated in the verification phase, the expected changes in computing and memory resource occupancy introduced, and at the same time combines a set of predefined quantitative deployment risk assessment inputs, which specifically include the predicted level of system stability after the introduction of the new version, the estimated probability and amplitude of regression of key performance indicators, the expected numerical value of the increased computing resource load, and the evaluation results of compatibility issues with existing system components. Through its internal established multi-factor weighted scoring logic, the deployment evaluation model comprehensively calculates and analyzes the above multi-dimensional input information to output a set of structured and quantitative deployment condition evaluation indicators. This indicator set clearly gives the expected comprehensive benefit score for deploying this new version of the optimization result, the estimated deployment success probability, and the level evaluation results of the accompanying specific risks.

[0064] Subsequently, the deployment strategy decision-making unit inside the optimization module receives and processes this set of deployment condition evaluation indicators output by the quantitative deployment evaluation model, which includes the expected comprehensive benefit score, the estimated success probability, and the risk levels of the new version. The core function of this decision-making unit is to perform logical judgments based on a pre-configured and structured deployment decision table. The internal of this decision table solidifies the rules for mapping the numerical intervals of different deployment condition evaluation indicators clearly and directly to specific predefined deployment action instructions. These rule entries are preset by the system after making a strategic balance among deployment benefits, introduced risks, and the current system stability. By referring to this deployment decision table and matching the current evaluation indicators, the decision-making unit finally generates a specific optimization command, which designates one from including immediately and comprehensively replacing the old version, selectively conducting a small-scale experimental deployment of the new version and simultaneously monitoring its performance, and determining that the current conditions are not suitable for deployment and suspending the operation. Finally, the optimization module executes the specified specific deployment operation according to this deployment action instruction, safely and effectively updates the optimization model and strategy into the corresponding functional modules in the system, and makes them take effect immediately in subsequent tasks.

[0065] By continuously deep learning and intelligently optimizing the data and quality indicators in the entire printing process, this optimization module can continuously improve the prediction accuracy of the core predictive control model in the system and the regulation efficiency of the key adaptive control strategy. This directly overcomes the problems of decreased and ineffective precise control ability caused by model aging, non-optimal parameters and strategies. Through strict verification and intelligent, risk-controlled automatic deployment of the optimized model and strategy, it ensures that the entire microcapsule printing system can continuously achieve and stably maintain a highly precise microcapsule image color development effect in the face of material property changes, equipment state evolution, and complex task requirements.

[0066] An intelligent control system for microcapsule images for printing provided by the present invention. The input module intelligently calculates and refines the initial control parameters through gradient boosting decision trees and optimization algorithms, laying a foundation for high-precision operations; the execution module combines online sensor feedback and parameter self-tuning PID control to achieve dynamic and precise deposition of microcapsules; the intelligent control module uses machine vision for real-time monitoring and dynamically optimizes the color development process through an adaptive control algorithm to ensure the accurate achievement of the final color characteristics. The optimization module continuously improves the long-term effectiveness and adaptability of the overall precise control of the system through learning and evaluation of the full-process data and model updates. This full-closed-loop intelligent design from initial setting, precise execution, real-time color development regulation to long-term optimization realizes the adaptive precision control of the entire process of microcapsule image construction.

[0067] Embodiment 2: Company B introduced an intelligent control method for microcapsule images for printing provided by the present invention in order to achieve intelligent and precise control of the penetration rate of the microcapsule image developer. The method flow chart is as Figure 3 shown, and the specific process is as follows: Receive the input digital image file as the basic program setting, integrate the inherent color gamut ability of the selected material and historical production data, intelligently calculate the control parameters, and generate execution commands and control signals; First, the system receives a digital design file containing target image information and the preset ideal developer penetration characteristic indicators for each region of the image. The penetration characteristic indicators are determined as the target penetration depth, penetration rate curve, and penetration uniformity standard, which are preset based on the correlation between the known penetration behavior of the selected microcapsules, developer, and substrate material system and the final color development effect. Subsequently, integrating the information in the penetration-related characteristic database of the selected materials (including the kinetic parameters of the influence of temperature on penetration) and the penetration and color development data in historical production, a Gradient Boosting Decision Tree (GBDT) model is used to process the above information to generate a set of preliminary temperature curve setting values aimed at achieving the target penetration characteristics and the initial parameters of the developer penetration kinetic model. Then, the Adaptive Perturbation Exploration Optimization (APEO) algorithm is used to refine and optimize the preliminary temperature curve setting values and model parameters, and finally output the initial temperature control program and control signal.

[0068] Receive the execution command and image data, combine with the real-time status information monitored by the online sensor, dynamically adjust the key process operation parameters, and output the substrate with the microcapsule deposition regulated according to the instruction and the relevant execution data.

[0069] Further, according to the deposition instruction included in the initial temperature control program generated in the previous step, the microcapsule material is deposited in the specified area of the substrate with high precision. During this process, the system uses the online sensor to monitor the status of the deposition process in real time, and dynamically adjusts the key process operation parameters through the parameter self-tuning PID control algorithm to ensure the accuracy of the deposition pattern, laying a foundation for subsequent precise penetration control.

[0070] Receive the substrate after deposition, control parameters, and execution command, monitor the color development characteristics in real time according to the program through the machine vision system, judge the deviation between the real-time monitored color development characteristics and the target color characteristics defined by the control parameters through the adaptive control algorithm, dynamically optimize and adjust the control parameters of the separable regions, and finally output the substrate with the color development completed and the detailed execution data; Further, after the microcapsule deposition is completed, it enters the stage of developer penetration regulation and color development activation. At this time, the system drives the precision temperature control component configured in the excitation unit to apply the starting temperature set by the initial temperature regulation program to the microcapsule area on the printing substrate. The developer penetration kinetics model built in the intelligent control module is activated. This model takes the region temperature, action time, and material properties monitored in real time as inputs, and dynamically predicts and outputs the current penetration rate of the developer and the achieved penetration depth. The adaptive control algorithm of the system receives the real-time penetration state data output by the developer penetration kinetics model, continuously compares it with the set target penetration characteristic index, and determines the deviation between the current penetration state and the target. Based on this deviation, combined with the monitoring results of the color change trend by the machine vision system, the adaptive control algorithm predicts the comprehensive impact of different temperature adjustment strategies on the subsequent penetration rate and final color development through the dynamic process response model, and then calculates and outputs precise adjustment instructions for the temperature control component to dynamically optimize the heating power, heating and cooling rate, and heat preservation duration. This adjustment aims to precisely guide the actual penetration process of the developer to the target trajectory, and at the same time coordinately control other necessary color development excitation parameters to ensure that the final image achieves the expected color characteristics.

[0071] Collect the execution data and quality indicators of each module in the whole process for evaluation to obtain the optimization command, and deploy and update the modules in the system according to the optimization command.

[0072] Further, collect the execution data of the whole process. In this embodiment, it particularly includes: the actual temperature curve applied, the prediction sequence of the penetration kinetics model, the color evolution data monitored by the machine vision, and the quantitative evaluation result of the improvement in color development quality brought by penetration optimization of the final product. Through in-depth analysis of this data, using the genetic algorithm, optimize the strategies of the GBDT model and APEO algorithm used to generate the initial temperature regulation program in the input module, and continuously learn and refine the parameters of the developer penetration kinetics model in the intelligent control module, as well as the dynamic process response model and heuristic rule set related to temperature-regulated penetration. According to the optimization command, deploy and update the relevant models and strategies in the system to achieve iterative improvement in the accuracy and effectiveness of the entire method in terms of intelligent control.

[0073] To verify the effectiveness of the method of the present invention, the original method of Company B and the method of the present invention were compared in terms of the color difference between the final product and the target, the average absolute error of the penetration depth, and the product rejection rate caused by poor penetration. Among them, the smaller the color difference between the final product and the target, the closer the color is to the target; the smaller the average absolute error of the penetration depth, the more precise the penetration control and the more reliable the function implementation; and the smaller the product rejection rate caused by poor penetration, the higher the qualified rate of the produced products. The same input data was selected for testing with both methods, and the results are shown in Table 2.

[0074] Table 2 Comparison Results between the Original Method of B and the Method of the Present Invention

[0075] The method of the present invention realizes precise control over each key link in the entire process of microcapsule application. It can precisely calculate the initial process parameters, accurately execute material deposition, and perform real-time adaptive adjustment on the dynamic color development and activation processes. Through continuous learning and iterative optimization, the system continuously improves the accuracy of its control strategy and the adaptability to working condition changes, thereby achieving comprehensive and high-level precise control.

[0076] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent control system for microcapsule images in printing, comprising, characterized in that: Input module: Receives the input digital image file as the basic program setting, integrates the inherent characteristic parameters of the selected material and historical operation data, intelligently calculates the control parameters, and generates execution commands and control signals; Execution module: Receives the execution command and image setting data, combines the real-time status information monitored by the online sensor, dynamically adjusts the key process operation parameters, and outputs the printing substrate for regulating the microcapsule deposition according to the instruction and related execution data; Intelligent control module: Receives the printing substrate after deposition processing, target characteristic parameters and execution commands, and through the online monitoring system, monitors the color development characteristics output in real time according to the program, judges the deviation between the color development characteristics monitored in real time and the target requirements defined by the target characteristic parameters through the adaptive control algorithm, dynamically optimizes and adjusts the excitation parameters of the separable area, and finally outputs the printing substrate that meets the set requirements and detailed execution data; Optimization module: Collects the execution data and performance indicators of the whole process of each module for evaluation, obtains the optimization command, and deploys and updates the modules in the system according to the optimization command.

2. The intelligent control system for microcapsule images used in printing according to claim 1, characterized in that, The specific acquisition process of the intelligent calculation control parameters is as follows: Process the basic program setting, the integrated solid color domain ability data of the selected material, and historical production data through the gradient boosting decision tree model to generate preliminary control parameters; take the preliminary control parameters as the initial solution, and judge according to the color development quality optimization target and process feasibility constraints. If no further optimization is required, the preliminary control parameters are used as the final control parameters after verification and correction; if further optimization is required, the adaptive perturbation exploration optimization algorithm is used for iterative refinement, and then the final control parameters are output.

3. The intelligent control system for microcapsule images in printing according to claim 2, wherein, The adaptive perturbation exploration optimization algorithm includes intelligently selecting the parameters to be adjusted and the perturbation direction based on the heuristic rule set; dynamically and adaptively setting the perturbation amplitude of the control parameters, iteratively generating candidate parameter combinations, and updating the optimal parameter set after verification and evaluation until the termination condition is met.

4. An intelligent control system for microcapsule images in printing according to claim 1, characterized in that, The dynamic adjustment of the key process operation parameters includes: continuously obtaining and parsing the feedback data reflecting the real-time application state of the current microcapsule material from the online sensor; immediately comparing the obtained real-time application state feedback data with the expected deposition target state received from the control instruction to determine the deviation between the actual application state and the expected deposition target state; According to the deviation, calculate the adjustment amount required for the key deposition parameters in real time, automatically apply this adjustment amount to the corresponding key deposition parameters, and continuously reduce the deviation, so that the actual application state of the microcapsule material dynamically approaches and maintains the expected deposition target state.

5. The intelligent control system for microcapsule images used in printing according to claim 4, wherein, The real-time calculation of the adjustment amount required for the key deposition parameters includes: Analyze the historical adjustment instruction data of the key deposition parameters and the corresponding real-time material application state feedback data provided by the online sensor, and use the recursive estimation algorithm to continuously estimate and update the key dynamic characteristic parameters of the microcapsule deposition process; Based on the key dynamic characteristic parameters, automatically and periodically calculate and set the proportional, integral, and differential gain parameters of the PID control algorithm according to the preset PID parameter tuning rules.

6. The intelligent control system for microcapsule images used in printing according to claim 1, characterized in that, The process of real-time monitoring the color development characteristics output by the online monitoring system according to the program includes: During the color development process, the image data of each separable area on the printing substrate during the color development process is captured in real time and processed to locate the target area and calculate the quantitative color parameters and color uniformity therefrom, which are used as the color development characteristics of the real-time monitoring output and fed back to the adaptive control algorithm.

7. The intelligent control system for microcapsule images used in printing according to claim 6, wherein, The relevant control parameters of the separable area are dynamically optimized and adjusted by the adaptive control algorithm. The relevant control parameters are excitation parameters. The specific processing flow is as follows: For each separable area, the adaptive control algorithm calculates and generates a set of modified excitation unit parameters aiming to minimize the deviation quantization result and make the color development characteristics approach the target based on the deviation quantization result between the current color development characteristics of the separable area monitored by machine vision in real time and the preset target color development characteristics, combined with the dynamic process response model. Subsequently, the modified excitation unit parameters are applied to the excitation unit of the corresponding separable area to adjust its working state in real time.

8. An intelligent control system for microcapsule images used in printing according to claim 1, characterized in that, The evaluation of collecting the full-process execution data and quality indicators of each module includes: Using the deployment decision model to process the execution data and quality indicators, outputting the risk assessment indicators and deployment gains, and using the decision model combined with historical deployment cases to process the risk assessment indicators and deployment gains to obtain the optimization command.

9. An intelligent control method for microcapsule images in printing, using an intelligent control system for microcapsule images in printing as described in claim 1, characterized in that: Receiving the input digital image file as the basic program setting, integrating the inherent characteristic parameters of the selected material and historical operation data, intelligently calculating the relevant control parameters, generating the execution command and control signal; receiving the execution command and image setting data, combining with the real-time status information monitored by the online sensor, dynamically adjusting the key process operation parameters, and outputting the printing substrate and related execution data for regulating the microcapsule deposition according to the instruction. Receiving the printing substrate after deposition processing, the target characteristic parameters and the execution command, real-time monitoring the color development characteristics output by the online monitoring system according to the program, judging the deviation between the color development characteristics monitored in real time and the target requirements defined by the target characteristic parameters through the adaptive control algorithm, dynamically optimizing and adjusting the relevant control parameters of the separable area, and finally outputting the printing substrate and detailed execution data that meet the set requirements; collecting the full-process execution data and performance indicators of each module for evaluation to obtain the optimization command, and performing deployment updates on the modules in the system according to the optimization command.

Images