Intelligent color matching method for powder coating based on artificial intelligence

Abstract

The invention is suitable for the technical field of artificial intelligence application, and provides an artificial intelligence-based powder coating intelligent color matching method, which comprises the following steps of: firstly obtaining target color information, and then generating color formula information according to the target color information based on a trained target color matching model. According to the method, an end-to-end nonlinear mapping relation directly from'formula data 'to'color data (LAB)' is put forward and constructed, the complexity of interaction of pigments and other materials in actual production can be better reflected, the interaction is completed in a virtual formula space, the number of physical experiments is greatly reduced, the continuous optimization and self-learning capabilities are achieved, and the method is suitable for industrial production. Dependence on specific experts is greatly reduced, and precipitation of color matching knowledge and standardization of the color matching process are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence application, in particular to a powder coating intelligent color matching method based on artificial intelligence. BACKGROUND

[0002] As an environmentally friendly, high utilization and excellent performance coating material, powder coating has been widely used in the fields of building materials, home appliances, furniture, automobiles, industrial equipment, etc. With the increasing market competition and the continuous improvement of consumer individualization demand, the market has put forward very high requirements for the diversity, accuracy and delivery speed of powder coating color. Therefore, how to quickly, accurately and low-costly match the target color has become one of the core problems to be solved in the powder coating industry.

[0003] The essence of color is the physical phenomenon perceived by the human eye or color measuring instrument after the interaction of light and object. The color matching of powder coating is to mix different kinds and proportions of base color pigments (such as titanium white, carbon black, phthalocyanine blue, permanent yellow, etc.) with resin, additives, etc., and after extrusion, tabletting, grinding and other processes, a coating layer that matches the target color sample in vision is prepared.

[0004] At present, the traditional color matching process highly depends on the personal experience and skills of color matching technicians. The color matching technician needs to preliminarily formulate a formula subjectively, then makes a sample according to the assumed formula, and then judges the direction of color deviation (such as red, green, light or dark) by experience, then manually adjusts the type or proportion of the corresponding pigment in the formula, and then performs a new round of trial and error. This process is repeated until the color difference meets the customer's requirements. The whole process is complicated, long in cycle, and high in trial and error cost, which seriously restricts the production efficiency and market response ability of enterprises, and has the problem of low reliability, which needs to be further improved. SUMMARY

[0005] Therefore, the embodiments of the present application provide a powder coating intelligent color matching method based on artificial intelligence to solve the problem of low reliability in the prior art.

[0006] In a first aspect, the embodiments of the present application provide a powder coating intelligent color matching method based on artificial intelligence, which comprises: obtaining target color information; generating color formula information according to the target color information based on a trained target color matching model.

[0007] Compared with the prior art, the beneficial effects are that the powder coating intelligent color matching method based on artificial intelligence provided by the embodiment of the present application can obtain target color information first, and then generate color formula information based on the trained target color matching model according to the target color information, thereby reducing the color matching process and reducing the trial and error cost, greatly improving the reliability, and to some extent solving the problem of low reliability at present.

[0008] In a second aspect, the embodiment of the present application provides a powder coating intelligent color matching system based on artificial intelligence, the system comprises: a target color information acquisition module, configured to acquire target color information; a color formula information generation module, configured to generate color formula information based on the trained target color matching model according to the target color information.

[0009] In a third aspect, the embodiment of the present application provides a terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the method of the first aspect when executing the computer program.

[0010] In a fourth aspect, the embodiment of the present application provides a computer readable storage medium, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the method of the first aspect.

[0011] It can be understood that the beneficial effects of the second aspect to the fourth aspect can be referred to the related description in the first aspect, which will not be repeated here. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiments or prior art description will be briefly introduced as follows.

[0013] Figure 1 is a flowchart of the powder coating intelligent color matching method provided by an embodiment of the present application; Figure 2 is a flowchart of the powder coating intelligent color matching method provided by an embodiment of the present application before step S200; Figure 3 is a first schematic diagram of the target color matching model provided by an embodiment of the present application; Figure 4 is a flowchart of step S102 in the powder coating intelligent color matching method provided by an embodiment of the present application; Figure 5 is a flowchart of step S200 in the powder coating intelligent color matching method provided by an embodiment of the present application; Figure 6 is a second schematic diagram of a target color matching model provided by an embodiment of the present application; Figure 7 is a flowchart diagram of a process after step S200 in the intelligent color matching method for powder coatings provided by an embodiment of the present application; Figure 8 is a third schematic diagram of a target color matching model provided by an embodiment of the present application; Figure 9 is a block diagram of an intelligent color matching system for powder coatings provided by an embodiment of the present application; Figure 10 is a schematic diagram of a terminal device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0014] In the following description, for the purpose of explanation and not limitation, specific details are set forth, such as particular system configurations, techniques, etc., in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

[0015] In the description of the present application and the appended claims, the terms "first", "second", "third", etc. are used only to distinguish descriptions, and cannot be understood as indicating or implying relative importance.

[0016] In the present application, the reference "one embodiment" or "some embodiments" means that the specific features, structures or characteristics described in connection with the embodiment are included in one or more embodiments of the present application. Therefore, the statements "in one embodiment", "in some embodiments", "in other some embodiments", "in further some embodiments" and the like appearing in different places in the present specification are not necessarily all referring to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized. The terms "include", "contain", "have" and their variants mean "include but not limited to", unless otherwise specifically emphasized.

[0017] In order to illustrate the technical solutions described in the present application, the following will be described by specific embodiments.

[0018] Please refer to Figure 1 , Figure 1is a flowchart of an intelligent powder coating color matching method based on artificial intelligence provided by the embodiment of the present application. In the present embodiment, the execution subject of the intelligent powder coating color matching method is a terminal device. It can be understood that the types of the terminal device include but are not limited to tablet computers, notebook computers, ultra-mobile personal computers (UMPC), netbooks, personal digital assistants (PDA), etc., and the specific type of the terminal device is not limited in the present embodiment.

[0019] At present, the traditional manual experience color matching method completely depends on the experience and subjective judgment of the color matching person, that is, the color matching person directly gives an initial formula by observing the target color sample with naked eyes and relying on the knowledge about pigment characteristics, hiding power, compatibility, cost and color superposition effect accumulated for a long time, and then gradually approaches the target color through the cycle of “trial and error-adjustment”. This method has the following disadvantages: (1) the color matching effect is strongly related to the technical level and state of the color matching person, and the knowledge is difficult to be deposited and standardized, and once the experienced color matching person leaves, it will cause great loss to the enterprise; (2) different color matching persons may have completely different understanding and formula adjustment strategies for the same color, resulting in unstable formula results; (3) a complex color, such as a metallic effect color or a high saturation color, may need dozens or even hundreds of times of sample printing and debugging, and the cycle is as long as several days or even weeks; (4) each sample printing needs to consume pigments, resins, electric energy and labor cost, and the cumulative cost is very huge.

[0020] Referring to Figure 1 The intelligent powder coating color matching method provided by the present embodiment includes but is not limited to the following steps: In S100, target color information is acquired.

[0021] Specifically, the terminal device can acquire the target color information, wherein the target color information can be in the form of LAB value.

[0022] In some possible implementation manners, in order to facilitate subsequent implementation of calculating accurate color formula, referring to Figure 2 Before step S200, the method further includes but is not limited to the following steps: In S101, historical data set information and an untrained target color matching model are acquired.

[0023] Specifically, the terminal device can acquire the historical data set information and the untrained target color matching model, wherein the historical data set information includes a plurality of historical data information, the number of historical data information can exceed 100,000 groups, each historical data information includes historical direction vector information and LAB vector value information, and the historical data set information can be pre-stored in a historical database.

[0024] In S102, the historical data set information is preprocessed to generate preprocessed historical data set information.

[0025] Specifically, after the terminal device obtains the historical data set information, the terminal device can preprocess the historical data set information to generate preprocessed historical data set information, thereby preparing for subsequent model training.

[0026] In S103, the untrained target color matching model is trained based on the preprocessed historical data set information to generate a trained target color matching model.

[0027] Exemplarily, please refer to Figure 3 After the terminal device generates the preprocessed historical data set information, the terminal device can train the untrained target color matching model based on the preprocessed historical data set information to generate a trained target color matching model, which can discover and memorize the complex and nonlinear correspondence between the pigment combination in the formula and the final color LAB value. The target color matching model can be a nonlinear prediction model based on deep neural network (DNN), gradient boosting decision tree (GBDT) or random forest (Random Forest).

[0028] In some possible implementations, in order to implement preprocessing of data, please refer to Figure 4 Step S102 includes but is not limited to the following steps: In S1021, the plurality of historical data set information in the historical data set information is subjected to data cleaning processing to generate first intermediate data set information.

[0029] Specifically, the terminal device can perform data cleaning processing on the plurality of historical data set information in the historical data set information to generate first intermediate data set information, wherein the first intermediate data set information is used to describe the historical data set information after data cleaning processing.

[0030] In S1022, the first intermediate data set information is subjected to normalization processing to generate second intermediate data set information.

[0031] Specifically, after the terminal device generates the first intermediate data set information, the terminal device can perform normalization processing on the first intermediate data set information to generate second intermediate data set information, wherein the second intermediate data set information is used to describe the first intermediate data set information after normalization processing.

[0032] In S1023, the second intermediate data set information is subjected to feature extraction processing to generate preprocessed historical data set information.

[0033] Specifically, after the terminal device generates the second intermediate data set information, the terminal device can perform feature extraction processing on the second intermediate data set information to generate preprocessed historical data set information, where the preprocessed historical data set information is used to describe a data collection of the plurality of second intermediate data set information after the feature extraction processing, so as to realize cleaning, normalization, feature extraction and the like of the original data, and prepare for model training.

[0034] In S200, based on the trained target color matching model, color formula information is generated according to the target color information.

[0035] Specifically, after the terminal device obtains the target color information, the terminal device can generate color formula information based on the trained target color matching model according to the target color information, so as to convert the color matching problem into an optimization solving problem when a new target color is received, use the trained target color matching model to quickly search and reverse inference in a huge and virtual formula space, and directly predict a feasible formula that can best match the target LAB value, which can be directly used for production and proofing.

[0036] In a possible implementation, the terminal device can provide a graphical interface for the user to input the target color, so that the user can view the predicted formula and the simulation effect.

[0037] In some possible implementations, in order to implement the generation of the color formula information, refer to Figure 5 and Figure 6 , the step S200 includes but is not limited to the following steps: In S210, the target color information is input to the trained target color matching model to generate one or more candidate formula information.

[0038] Specifically, the terminal device can input the target color information to the trained target color matching model to generate one or more candidate formula information, which is the optimal formula in Figure 6 .

[0039] In S220, in response to an experimental proofing completion instruction, actual LAB value information corresponding to each candidate formula information is obtained.

[0040] Specifically, after the terminal device generates one or more candidate formula information, the terminal device can obtain actual LAB value information corresponding to each candidate formula information in response to an experimental proofing completion instruction, where the experimental proofing completion instruction is used to instruct a test personnel to complete experimental proofing according to the predicted formula.

[0041] In S230, for each candidate formula information: actual color difference information is generated based on actual LAB value information and target color information.

[0042] Specifically, after the terminal device obtains the actual LAB value information, the terminal device can generate actual color difference information for each candidate formula based on the actual LAB value information and the target color information. The actual color difference information is used to describe the color difference between the actual LAB value information and the target color information.

[0043] In S240, it is determined whether the actual color difference information meets the preset acceptance standard.

[0044] Specifically, after the terminal device generates the actual color difference information, the terminal device can determine whether the actual color difference information meets the preset qualification standard. The qualification standard is used to describe whether the actual color difference information is greater than the minimum value of the preset allowable color difference range information and less than the minimum value of the allowable color difference range information.

[0045] In S250, if the actual color difference information meets the qualification standard, the candidate formula information that meets the qualification standard is determined as the color formula information. Otherwise, based on the candidate formula information and the actual LAB value information, a new historical data information is generated. Then, the target color matching model is trained again based on the new historical data information. Then, the target color information is input into the trained target color matching model again to generate one or more candidate formula information until it is determined whether the actual color difference information meets the preset qualification standard, until the actual color difference information meets the qualification standard.

[0046] Specifically, if the actual color difference information meets the qualified standard, the terminal device can determine the candidate formula information that meets the qualified standard as the color formula information; otherwise, based on the candidate formula information and the actual LAB value information, a new historical data information is generated, and the target color matching model is trained again based on the new historical data information. Then, the above steps S210 to S240 are executed again until the actual color difference information meets the qualified standard.

[0047] For example, as shown in Table 1 below, the applicant conducted a large number of experiments on the intelligent color matching method for powder coatings according to the embodiments of this application; the experimental results show that the intelligent color matching method for powder coatings can not only efficiently generate a relatively accurate formula effect, but also significantly improve the accuracy and consistency of color matching, providing strong technical support for the application of powder coatings.

[0048] Table 1 Test Results of the Target Color Matching Model

[0049] In some possible implementations, during the model training phase of the target color matching model, the terminal device utilizes massive amounts of historical data to train the model, enabling it to learn the mapping from recipes to colors. During the prediction application phase of the target color matching model, the user first provides the standard LAB values ​​(L0, a0, b0) of the target color. Then, the terminal device transforms the user's requirement into an optimization problem: finding a recipe X such that the AI ​​model predicts LAB_pred = ... The process of minimizing the difference between Model(X) and the target (L0,a0,b0) is not a trial-and-error process in a laboratory, but rather a rapid process completed in a virtual recipe space on a computer using optimization algorithms (such as gradient descent, genetic algorithms, or particle swarm optimization). The target color matching model then outputs one or more feasible recipes with the smallest color difference, lowest cost, or most stable performance. During the verification and feedback loop phase, the terminal device can conduct a small number of experiments to verify the recipes recommended by the target color matching model, such as one or two verifications. Regardless of success or failure, the recipes and actual LAB values ​​from this experiment will be fed back to the historical database as new data samples. Through this feedback loop mechanism, the target color matching model can continuously learn, evolve, and become increasingly accurate.

[0050] For some possible implementations, please refer to [link / reference needed] to improve reliability. Figure 7 After step S200, the method further includes, but is not limited to, the following steps: In S201, another new historical data information is generated based on the color formula information and the actual LAB value information.

[0051] Specifically, the terminal device can generate another new historical data information based on the color formula information and the actual LAB value information.

[0052] In S202, the target color matching model is retrained based on another new historical data, generating a retrained target color matching model.

[0053] Specifically, after the terminal device generates another new historical data information, the terminal device can retrain the target color matching model based on the new historical data information to generate a retrained target color matching model.

[0054] In some possible implementations, the working principle of the target color model can be as follows: Figure 8As shown, the input layer of the target color matching model can represent a formula, with each node representing the normalized proportion of a primary color pigment or filler in the formula, forming a mathematical vector describing the formula. The hidden layer of the target color matching model is the core of the model. Through a multi-layer neuron structure and a non-linear activation function (such as ReLU), the network can learn the complex, non-linear superposition and interaction relationships between pigments. For example, it can learn deep-seated rules such as "when pigment A and pigment B are mixed in a certain proportion, it will lead to a significant increase in the b value (yellow-blue axis)". The output layer of the target color matching model can have three nodes, each corresponding to the predicted color LAB value. The training process of the target color matching model can be achieved by using a backpropagation algorithm to continuously adjust the connection weights between neurons in the network, minimizing the difference between the model's predicted output Y_pred and the true measured value Y_true, such as minimizing the loss function and the mean squared error (MSE).

[0055] In some possible implementations, the target color matching model can also consider factors such as cost, pigment inventory, or formula stability when outputting the predicted formula to perform multi-objective optimization and output the Pareto optimal solution set. The color measurement data in the verification and feedback loop stage can be the Lab value of the CIELAB color space, or the value of other color spaces (such as LCH, XYZ).

[0056] The implementation principle of the intelligent color matching method for powder coatings based on artificial intelligence in this application is as follows: The terminal device can first acquire the target color information, and then generate color formula information based on the target color information according to the trained target color matching model. This abandons the traditional KM theory model based on ideal assumptions and proposes and constructs for the first time an end-to-end nonlinear mapping relationship directly from "formula data" to "color data (LAB)". This relationship is "learned" from massive historical data by machine learning / deep learning models, rather than "calculated" by physical formulas. It better reflects the complexity of pigment interactions in actual production and solves the classic "reverse engineering" problem in the field of color matching. Using a trained AI model, it reverse-engineers and quickly searches for "feasible formulas" from "target color (LAB value)". This is an optimized solution process completed in a virtual formula space, which greatly reduces the number of physical experiments. It also features a continuous feedback loop. Regardless of success or failure, each experimental verification result of the predicted formula is fed back to the historical database as a new data sample for iterative updates and retraining of the model. This gives the system the ability to continuously optimize and learn, becoming more accurate with use and forming a constantly evolving intelligent system. At the same time, it transforms the "experience" and "feeling" that originally existed in the minds of colorists and were difficult to quantify into explicit algorithmic models that can be stored, replicated, and optimized through the target color matching model. This reduces the dependence on specific experts and realizes the accumulation of color matching knowledge and the standardization of the color matching process.

[0057] It should be noted that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0058] Embodiments of this application also provide an intelligent color matching system for powder coatings based on artificial intelligence. For ease of explanation, only the parts relevant to this application are shown, such as... Figure 8 As shown, the system 80 includes: Target color information acquisition module 81: Used to acquire target color information; Color recipe information generation module 82: Used to generate color recipe information based on the target color information and the trained target color matching model.

[0059] Optionally, the system 80 also includes: Historical dataset information acquisition module: used to acquire historical dataset information and untrained target color matching model. The historical dataset information includes multiple historical data information, each of which includes historical recipe vector information and LAB vector value information. Historical dataset information generation module: used to preprocess historical dataset information and generate preprocessed historical dataset information; Target color matching model generation module: This module is used to train an untrained target color matching model based on preprocessed historical dataset information, and generate a trained target color matching model.

[0060] Optionally, the above-mentioned historical dataset information generation module includes: First intermediate dataset information generation submodule: Used to perform data cleaning and processing on multiple historical datasets to generate first intermediate dataset information; The second intermediate dataset information generation submodule is used to normalize the information of the first intermediate dataset and generate the information of the second intermediate dataset. Historical dataset information generation submodule: used to perform feature extraction processing on the second intermediate dataset information to generate preprocessed historical dataset information.

[0061] Optionally, the color formula information generation module 82 mentioned above includes: Candidate recipe information generation submodule: Used to input target color information into the trained target color matching model and generate one or more candidate recipe information; Actual LAB value information acquisition submodule: In response to the experimental sampling completion command, it acquires the actual LAB value information corresponding to each candidate formulation; Actual color difference information generation submodule: used to generate actual color difference information for each candidate formula based on actual LAB value information and target color information; Actual color difference information judgment submodule: used to determine whether the actual color difference information meets the preset qualified standard. The qualified standard is used to describe whether the actual color difference information is greater than the minimum value of the preset allowable color difference range information and less than the minimum value of the allowable color difference range information. The color formula information determination submodule is used to determine the candidate formula information that meets the qualified standard if the actual color difference information meets the qualified standard. Otherwise, it generates a new historical data information based on the candidate formula information and the actual LAB value information. Then, it retrains the target color matching model based on the new historical data information. Then, it re-executes the process of inputting the target color information into the trained target color matching model to generate one or more candidate formula information until it judges whether the actual color difference information meets the preset qualified standard, until the actual color difference information meets the qualified standard.

[0062] Optionally, if the actual color difference information meets the acceptable standard, the system 80 also includes: Historical data information generation module: used to generate another new historical data information based on color formula information and actual LAB value information; Target color matching model generation module: used to retrain the target color matching model based on new historical data, and generate a retrained target color matching model.

[0063] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0064] This application also provides a terminal device, such as... Figure 9 As shown, the terminal device 90 in this embodiment includes: a processor 91, a memory 92, and a computer program 93 stored in the memory 92 and executable on the processor 91. When the processor 91 executes the computer program 93, it implements the steps in the above-described embodiment of the intelligent color matching method for powder coatings, for example... Figure 1 Steps S100 to S200 are shown; or, when processor 91 executes computer program 93, it implements the functions of each module in the above-described device, for example... Figure 8 The functions of modules 81 to 82 are shown.

[0065] The terminal device 90 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. The terminal device 90 includes, but is not limited to, a processor 91 and a memory 92. Those skilled in the art will understand that... Figure 9 This is merely an example of terminal device 90 and does not constitute a limitation on terminal device 90. It may include more or fewer components than shown, or combine certain components, or different components. For example, terminal device 90 may also include input / output devices, network access devices, buses, etc.

[0066] The processor 91 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.; the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0067] The memory 92 can be an internal storage unit of the terminal device 90, such as the hard disk or memory of the terminal device 90. The memory 92 can also be an external storage device of the terminal device 90, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the terminal device 90. Furthermore, the memory 92 can include both internal storage units and external storage devices of the terminal device 90. The memory 92 can also store computer program 93 and other programs and data required by the terminal device 90. The memory 92 can also be used to temporarily store data that has been output or will be output.

[0068] One embodiment of this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0069] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the methods, principles and structures of this application should be covered within the scope of protection of this application.

Claims

1. An intelligent color matching method for powder coatings based on artificial intelligence, characterized in that, The method includes: Obtain target color information; Based on the trained target color matching model, color formula information is generated according to the target color information.

2. The method according to claim 1, characterized in that, Before generating color recipe information based on the target color information using the trained target color matching model, the method further includes: Obtain historical dataset information and an untrained target color matching model, wherein the historical dataset information includes multiple historical data information, and each of the historical data information includes historical recipe vector information and LAB vector value information; The historical dataset information is preprocessed to generate preprocessed historical dataset information. Based on the preprocessed historical dataset information, the untrained target color matching model is trained to generate a trained target color matching model.

3. The method according to claim 2, characterized in that, The process of preprocessing historical dataset information to generate preprocessed historical dataset information includes: Data cleaning is performed on multiple historical datasets to generate the first intermediate dataset. The information in the first intermediate dataset is normalized to generate the information in the second intermediate dataset. Feature extraction is performed on the second intermediate dataset information to generate preprocessed historical dataset information.

4. The method according to claim 1, characterized in that, The training-based target color matching model generates color formula information based on the target color information, including: The target color information is input into the trained target color matching model to generate one or more candidate recipes. In response to the experimental sampling completion command, the actual LAB value information corresponding to each of the candidate formulations is obtained; For each of the candidate formulations: generate actual color difference information based on the actual LAB value information and the target color information; Determine whether the actual color difference information meets the preset qualification standard, wherein the qualification standard is used to describe that the actual color difference information is greater than the minimum value of the preset allowable color difference range information and less than the minimum value of the allowable color difference range information; If the actual color difference information meets the qualification standard, then the candidate formula information that meets the qualification standard is determined as color formula information; otherwise, based on the candidate formula information and the actual LAB value information, a new historical data information is generated, and then the target color matching model is trained again based on the new historical data information. Then, the process of inputting the target color information into the trained target color matching model is executed again to generate one or more candidate formula information until the actual color difference information meets the preset qualification standard, until the actual color difference information meets the qualification standard.

5. The method according to claim 4, characterized in that, If the actual color difference information meets the qualification standard, then after generating color formula information based on the target color information according to the trained target color matching model, the method further includes: Based on the color formula information and the actual LAB value information, another new historical data information is generated; The target color matching model is retrained based on the new historical data to generate a retrained target color matching model.

6. An intelligent color matching system for powder coatings based on artificial intelligence, characterized in that, The system includes: Target color information acquisition module: used to acquire target color information; Color recipe information generation module: used to generate color recipe information based on the target color information, according to the trained target color matching model.

7. The system according to claim 6, characterized in that, The system also includes: Historical dataset information acquisition module: used to acquire historical dataset information and untrained target color matching model, wherein the historical dataset information includes multiple historical data information, and each of the historical data information includes historical recipe vector information and LAB vector value information; Historical dataset information generation module: used to preprocess historical dataset information and generate preprocessed historical dataset information; Target color matching model generation module: used to train the untrained target color matching model based on the preprocessed historical dataset information, and generate a trained target color matching model.

8. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 5.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images