Color matching recommendation method and system

Abstract

The invention provides a color matching recommendation method and system, and belongs to the field of textile printing and dyeing, and the method comprises the steps: building a color matching formula historical database; obtaining a color matching data set based on a color matching formula historical database and a decision tree classification model; training a pre-constructed LightGBM initial classification model on the basis of the color matching data set to obtain a LightGBM classification model; acquiring sample color parameters, sample material parameters and sample process parameters according to the sample to be subjected to color matching; obtaining a target formula database subset from a color matching formula historical database based on the sample material parameters and the sample process parameters; sample color parameters are input into the Light GBM classification model, so that the Light GBM classification model outputs a sample color system corresponding to a sample to be subjected to color matching; obtaining a target color system subset from the target formula database subset based on the sample color system; and according to the sample color parameters and the target color system subset, obtaining a color matching scheme meeting a preset requirement. According to the method, the color matching process can be intelligentized, and the accuracy and the recommendation efficiency of color matching formula recommendation are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of textile printing and dyeing industry, and in particular to a color matching recommendation method and system. BACKGROUND

[0002] Color matching technology is an important production link in the textile printing and dyeing industry. The existing color matching methods mainly rely on manual experience, which is inefficient. By now, the existing mainstream technologies mainly have the following problems: if the manual experience formula management is used, the formula management is not standardized, because it mainly relies on handwriting or paper records, which is difficult to query and save. At the same time, the formula efficiency is excessively dependent on manual experience, resulting in low efficiency of novice technicians.

[0003] Although the existing technology has an intelligent color matching system, the existing intelligent color matching system lacks relevant research on the establishment of a formula history database and database matching, which can easily lead to a lack of pertinence in recommendation, low efficiency, and an unclear matching result determination process. For example, if there is no formula with the same color difference in the historical data, the formula result cannot be output. Or the formula concentration needs to be revised in combination with the actual color matching process, which is inefficient and greatly depends on manual experience. SUMMARY

[0004] The present application aims to provide a color matching recommendation method and system to solve the above technical problems and improve the accuracy and efficiency of color matching formula recommendation.

[0005] To solve the above technical problems, the present application provides a color matching recommendation method, comprising the following steps:

[0006] establishing a color matching formula history database;

[0007] obtaining a plurality of color matching data sets based on the color matching formula history database and a preset decision tree classification model;

[0008] training a pre-constructed LightGBM initial classification model based on the plurality of color matching data sets to obtain a LightGBM classification model;

[0009] obtaining sample color value parameters, sample material parameters, and sample process parameters according to a sample to be color matched;

[0010] obtaining a target formula database subset from the color matching formula history database based on the sample material parameters and sample process parameters;

[0011] inputting the sample color value parameters into the LightGBM classification model to make the LightGBM classification model output a sample color system corresponding to the sample to be color matched;

[0012] acquire a target color scheme subset from the target formula database subset based on the sample color system;

[0013] acquire a color matching scheme meeting preset requirements according to the sample color value parameter and the target color scheme subset.

[0014] In the above scheme, first, the color matching formula history database is established by collecting existing color matching data, and the color matching data set is obtained by screening through the preset decision tree classification model. On this basis, the pre-constructed LightGBM initial classification model is trained to obtain the LightGBM classification model, and the color matching data set is further divided into a color scheme subset. For the sample to be color matched, first, the color value parameter, material parameter and process parameter thereof are acquired, the target formula database subset is screened out from the color matching formula history database according to the material parameter and process parameter, and then the color parameter is input into the LightGBM classification model to determine the color scheme, and the target color scheme subset is further acquired from the target formula database subset. Finally, the color matching scheme meeting the preset requirements is acquired according to the color parameter and the target color scheme subset.

[0015] The above scheme digitizes the dyeing formula data, and uses the decision tree classification model and the LightGBM classification model to efficiently group and screen the formula data, so as to intelligently and automatically realize the color matching process, reduce the dependence on manual experience, and also improve the accuracy and efficiency of color matching formula recommendation.

[0016] Further, the acquiring of the plurality of color matching data sets based on the color matching formula history database and the preset decision tree classification model comprises: extracting material characteristics and process characteristics of different color matching formulas from the color matching formula history database to obtain a material characteristic set and a process characteristic set; and inputting the material characteristic set and the process characteristic set into the decision tree classification model, so that the decision tree classification model splits the color matching formula history database into a plurality of color matching data sets based on the material characteristic set and the process characteristic set.

[0017] The color matching data set is a data set composed of color matching formula history data with the same material characteristics and the same process characteristics.

[0018] In the above scheme, in order to obtain a color matching data set composed of color matching formula history data with the same material characteristics and the same process characteristics, a decision tree classification model is used to filter in the color matching formula history database according to requirements such as fabric material and dyeing process parameters. The decision tree is a recursive tree structure, which classifies or regresses data through a series of rules. The decision tree classification model divides the data set into multiple subsets by selecting the optimal split point, so that the purity (or impurity) of each subset is maximally improved (or decreased). By presetting the filtering of the decision tree classification model, the color matching formula history database can be divided into several color matching data sets based on the material characteristic set and the process characteristic set for subsequent construction of the lightGBM model.

[0019] Further, the LightGBM classification model is obtained by training the pre-constructed LightGBM initial classification model based on the plurality of color matching data sets, including: performing data preprocessing based on the color matching formula history database, and dividing the data set into a training set and a test set; training the pre-constructed LightGBM initial classification model based on the training set and specified parameters to obtain a training model; performing performance evaluation and parameter optimization on the training model based on a preset cross-validation method to obtain a validation model; evaluating the validation model based on the test set to determine whether the validation model meets the preset training requirements, and obtaining the LightGBM classification model if the validation model meets the preset training requirements.

[0020] Further, if the validation model does not meet the preset training requirements, the validation model is retrained based on the training set and specified parameters to obtain a new training model and a new validation model, until the new validation model meets the preset training requirements.

[0021] In the above scheme, the color matching formula history database needs to be preprocessed first, including dividing the color matching formula history database formula color into 7 color systems (red, orange, yellow, green, blue, blue, and purple) according to the CIELAB color model and the OHTA color space, to establish a pre-constructed LightGBM initial classification model and divide the data set into a training set and a test set. Then, the pre-constructed LightGBM initial classification model is trained based on the training set and specified parameters to obtain a training model. In the training process, the performance of the training model is evaluated and the parameters are optimized based on a preset cross-validation method to obtain a validation model, and the validation model is evaluated based on the test set to determine whether the validation model meets the preset training requirements, and the LightGBM classification model is obtained if the validation model meets the preset training requirements.

[0022] It should be noted that if the verification model does not meet the preset index requirement, the training of the verification model should be performed again based on the training set and the specified parameters, thereby obtaining a new training model and a new verification model, until the new verification model meets the preset training requirement.

[0023] Further, the obtaining of the color matching scheme meeting the preset requirement according to the sample color value parameter and the target color system subset includes: respectively calculating the Euclidean distance between the sample color value parameter and each color value of the target color system subset to obtain a distance value sequence; and obtaining a target distance value meeting the preset requirement in the distance value sequence, and taking the target color system subset corresponding to the target distance value as the color matching scheme.

[0024] Further, the respectively calculating the Euclidean distance between the sample color value parameter and each color value of the target color system subset to obtain a distance value sequence specifically includes:

[0025] For any color system in the target color system subset, the Euclidean distance calculation process between the color value of the color system and the sample color value parameter is as follows:

[0026]

[0027] wherein ρ is the Euclidean distance, R x , G x , B x is the sample color value parameter, R n , G n , B n is the color value parameter of any color system in the target color system subset.

[0028] The distance value sequence is a data sequence composed of the distance values between each color value and the sample color value parameter.

[0029] The Euclidean distance obtained here is used to represent the color difference between the sample and each color system in the target color system subset. It should be noted that after the distance value sequence obtained by traversing the sample color value and the target color system subset is sorted from small to large, the dyeing formula corresponding to the first few distance values can be selected as the recommended formula output.

[0030] The application also provides a color matching recommendation system, comprising: a database construction module for establishing a color matching formula history database; a color matching classification module for obtaining a plurality of color matching data sets based on the color matching formula history database and a preset decision tree classification model; a classification model training module for training a pre-constructed LightGBM initial classification model based on the plurality of color matching data sets to obtain a LightGBM classification model; a sample parameter acquisition module for acquiring sample color value parameters, sample material parameters and sample process parameters according to a sample to be color matched; a target formula screening module for acquiring a target formula database subset from the color matching formula history database based on the sample material parameters and sample process parameters; a sample color system acquisition module for inputting the sample color value parameters into the LightGBM classification model to enable the LightGBM classification model to output a sample color system corresponding to the sample to be color matched; a target color system acquisition module for acquiring a target color system subset from the target formula database subset based on the sample color system; and a color matching recommendation module for acquiring a color matching scheme meeting a preset requirement according to the sample color value parameters and the target color system subset.

[0031] The system provided by the above scheme is simple to construct, and the existing formula history data is ingeniously used to construct a lightGBM classification model and used for formula recommendation for a sample to be color matched. The system efficiently groups and screens the formula data by using a decision tree classification model and a LightGBM classification model, and integrates the color matching recommendation method on a color matching recommendation system for use by a color matching operator, so that the color matching process is intelligentized and automated, the dependence of formula management and recommendation process on manual experience is reduced, and the accuracy and recommendation efficiency of color matching formula recommendation are improved.

[0032] Further, the color matching classification module is configured to obtain a plurality of color matching data sets based on the color matching formula history database and a preset decision tree classification model, comprising: extracting material characteristics and process characteristics of different color matching formulas from the color matching formula history database to obtain a material characteristic set and a process characteristic set; inputting the material characteristic set and the process characteristic set into the decision tree classification model, so that the decision tree classification model splits the color matching formula history database into a plurality of color matching data sets based on the material characteristic set and the process characteristic set; and the color matching data set is a data set composed of color matching formula history data with the same material characteristics and the same process characteristics.

[0033] Further, the classification model training module is configured to train a pre-constructed LightGBM initial classification model based on a plurality of color matching data sets to obtain a LightGBM classification model, including: performing data preprocessing based on the color matching formula history database, and dividing the data set into a training set and a test set; training the pre-constructed LightGBM initial classification model based on the training set and specified parameters to obtain a training model; performing performance evaluation and parameter optimization on the training model based on a preset cross-validation method to obtain a verification model; and evaluating the verification model based on the test set to determine whether the verification model meets preset training requirements, and obtaining the LightGBM classification model if the verification model meets the preset training requirements.

[0034] Further, the color matching recommendation module is configured to obtain a color matching scheme meeting preset requirements according to the sample color value parameter and the target color system subset, including: calculating the Euclidean distance between the sample color value parameter and the color value of each color system in the target color system subset respectively to obtain a distance value sequence; and obtaining a target distance value meeting preset requirements in the distance value sequence, and taking the target color system subset corresponding to the target distance value as the color matching scheme.

[0035] The above scheme is a specific implementation of the color matching recommendation method on the color matching recommendation system. Through the operation of the color matching classification, classification model training, color matching recommendation and other modules, the operator can accurately and quickly obtain a color matching recommendation scheme with the help of the existing color matching formula history data, so as to intelligently and automatically realize the color matching process and reduce the dependence on manual experience. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 A color matching recommendation method flowchart is provided for an embodiment of the present application.

[0037] Figure 2 A color matching recommendation system architecture diagram is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0039] Please refer to Figure 1 The present embodiment provides a color matching recommendation method, including the following steps:

[0040] Step S1: establishing a color matching formula history database;

[0041] Step S2: obtaining a plurality of color matching data sets based on the color matching formula history database and a preset decision tree classification model;

[0042] Step S3: training a pre-constructed LightGBM initial classification model based on the plurality of color matching data sets to obtain a LightGBM classification model;

[0043] Step S4: obtaining sample color value parameters, sample material parameters and sample process parameters according to a sample to be color matched;

[0044] Step S5: obtaining a target formula database subset from the color matching formula history database based on the sample material parameters and sample process parameters;

[0045] Step S6: inputting the sample color value parameters into the LightGBM classification model to make the LightGBM classification model output a sample color family corresponding to the sample to be color matched;

[0046] Step S7: obtaining a target color family subset from the target formula database subset based on the sample color family;

[0047] Step S8: obtaining a color matching scheme meeting a preset requirement according to the sample color value parameters and the target color family subset.

[0048] In this embodiment, firstly, color matching data is collected to establish a color matching formula history database, and a preset decision tree classification model is used for screening to obtain color matching data sets. The color matching data can include formula serial number, cloth material, dyeing process parameters (processing temperature, dye source, holding time), color value (RGB value), dye type and concentration, etc. On this basis, a pre-constructed LightGBM initial classification model is trained to obtain a LightGBM classification model, and the color matching data sets are further divided into color family subsets. For a sample to be color matched, its color value parameters, material parameters and process parameters are obtained first, and then the target formula database subset is screened from the color matching formula history database according to the material parameters and process parameters. Then, the color parameters are input into the LightGBM classification model to determine the color family, and then the target color family subset is obtained from the target formula database subset. Finally, the color matching scheme meeting the preset requirement is obtained according to the color parameters and the target color family subset.

[0049] In this embodiment, the dyeing formula data is digitized, and the decision tree classification model and the LightGBM classification model are used to efficiently group and screen the formula data, so as to intelligently and automatically match the color, reduce the dependence on manual experience, and improve the accuracy and efficiency of color matching formula recommendation.

[0050] Further, the obtaining a plurality of color matching data sets based on the color matching formula history database and a preset decision tree classification model comprises: extracting material characteristics and process characteristics of different color matching formulas from the color matching formula history database to obtain a material characteristic set and a process characteristic set; and inputting the material characteristic set and the process characteristic set into the decision tree classification model, so that the decision tree classification model splits the color matching formula history database into a plurality of color matching data sets S based on the material characteristic set and the process characteristic set. N (N = 1, 2, 3, …);

[0051] The color matching data set is a data set composed of color matching formula historical data of the same material characteristics and the same process characteristics.

[0052] In the embodiment, in order to obtain the color matching data set composed of color matching formula historical data of the same material characteristics and the same process characteristics, a decision tree classification model is used to filter in the color matching formula history database according to requirements, such as for cloth material and dyeing process parameters. The decision tree is a recursive tree structure, which classifies or regresses data through a series of rules. The decision tree classification model splits the color matching formula history database into a plurality of color matching data sets by selecting the optimal split point, so that the purity (or impurity) of each color matching data set is maximally improved (or decreased). The key principles of the construction include the following three parts:

[0053] (1) Information gain: used for classification tasks, the information gain of the data set after being segmented by a certain feature is calculated, and the feature with the maximum information gain is selected for splitting.

[0054]

[0055] wherein D is a data set, A is a feature, v is a value of the feature A, and D v is a data subset with the value v of the feature A.

[0056] (2) Entropy: measures the purity of the data set.

[0057]

[0058] wherein p i is the proportion of class i in the data set.

[0059] (3) Gini Index: used for classification tasks, the optimal split point is selected by calculating the Gini Index.

[0060]

[0061] The color matching formula history database can be divided into several color matching data sets based on the material feature set and the process feature set through screening of a preset decision tree classification model for subsequent construction of a lightGBM model.

[0062] Further, the pre-constructed LightGBM initial classification model is trained based on the several color matching data sets to obtain a LightGBM classification model, including: data preprocessing based on the color matching formula history database, and dividing the data set into a training set and a test set; training the pre-constructed LightGBM initial classification model based on the training set and specified parameters to obtain a training model; performing performance evaluation and parameter optimization on the training model based on a preset cross-validation method to obtain a verification model; evaluating the verification model based on the test set to determine whether the verification model meets preset training requirements, and obtaining the LightGBM classification model if the verification model meets the preset training requirements.

[0063] Further, if the verification model does not meet the preset training requirements, the verification model is retrained based on the training set and specified parameters to obtain a new training model and a new verification model, until the new verification model meets the preset training requirements.

[0064] This embodiment uses LightGBM to construct a model for color system classification. LightGBM is a distributed gradient boosting framework based on decision tree algorithm, which is efficient and accurate, and is suitable for classification and regression tasks of large-scale data. LightGBM adopts a histogram-based decision tree learning algorithm. It constructs a histogram by dividing continuous data into buckets to speed up the calculation and reduce memory consumption. Unlike traditional layer-by-layer growth, LightGBM adopts a leaf node growth strategy, which selects the optimal leaf node for growth at each split to reduce the loss function faster. LightGBM discretizes continuous feature values into histograms to reduce the computational complexity of feature splitting, and quickly finds the optimal split point using the cumulative properties of histograms. Among them, by constructing a histogram, continuous feature values are divided into multiple buckets, each bucket contains a range of continuous values as follows:

[0065]

[0066] where H j is the cumulative value of the jth bucket, and I is an indicator function.

[0067] The goal of LightGBM is to optimize the following loss function by fitting the residual step by step:

[0068]

[0069] Where y is the actual value, F(x) is the predicted value, and l is the loss function for a single sample.

[0070] At the same time, LightGBM supports multi-threading and GPU acceleration, and realizes efficient model training through distributed learning. LightGBM can also handle large-scale categorical features, avoiding the high-dimensional problem caused by one-hot encoding in traditional methods.

[0071] In this embodiment, first, the color matching formula history database needs to be preprocessed, including dividing the color matching formula history database formula color into 7 color systems (red, orange, yellow, green, cyan, blue, purple) according to the CIELAB color model and OHTA color space, to establish a pre-constructed LightGBM initial classification model and divide the data set into a training set and a test set. Then, based on the training set and specifying parameters, the pre-constructed LightGBM initial classification model is trained to obtain a training model. In the training process, the performance of the training model also needs to be evaluated and the parameters are optimized based on the preset cross-validation method to obtain a validation model, and the validation model is evaluated based on the test set to determine whether the validation model meets the preset training requirements, and then the LightGBM classification model is obtained.

[0072] It should be noted that if the validation model does not meet the preset index requirements, the training of the validation model should be re-performed based on the training set and by specifying parameters, and then a new training model and a new validation model are obtained until the new validation model meets the preset training requirements.

[0073] It should be noted that the model performance needs to be evaluated by the cross-validation method during the model training process, and the hyperparameters need to be adjusted to optimize the model effect. The specific process is as follows:

[0074] (1) Data preparation: divide the data into features and labels, and use the 80 / 20 ratio to divide the data into a training set and a test set, and finally standardize the features to ensure that different features have the same scale. The corresponding encoding of the data label (color classification) is as follows:

[0075]

[0076]

[0077] (2) Define the model and the parameter grid: select the LightGBM classification model, and list the hyperparameters that need to be adjusted and their value ranges.

[0078] (3) Cross-validation and grid search: use GridSearchCV for hyperparameter search and cross-validation, run the grid search on the training set, and find the optimal parameter combination.

[0079] (4) Evaluate the optimal model: train the model using the optimal parameters: train the final model according to the optimal parameters found by grid search, and evaluate the model performance with the test set.

[0080] It should be noted that after the LightGBM classification model is constructed, the same material characteristics, the same process characteristics database is further divided into 7 color system subsets S NK (N = 1, 2, 3, …, K = 1, 2, … 7) according to the sample material, process characteristics, and the selected color matching data set S N is obtained after measuring the sample color value, and the target color system subset S NK is selected.

[0081] Further, the color matching scheme meeting the preset requirement is obtained according to the sample color value parameter and the target color system subset, comprising: calculating the Euclidean distance between the sample color value parameter and each color value in the target color system subset S NK to obtain a distance value sequence; obtaining a target distance value meeting the preset requirement in the distance value sequence, and taking the target color system subset corresponding to the target distance value as the color matching scheme.

[0082] Further, the Euclidean distance between the sample color value parameter and each color value in the target color system subset is calculated to obtain a distance value sequence, specifically:

[0083] For any color system in the target color system subset, the Euclidean distance calculation process of the color system color value and the sample color value parameter is:

[0084]

[0085] Where p is the Euclidean distance, R x , G x , B x is the sample color value parameter, R n , G n , B n is the color value parameter of any color system in the target color system subset.

[0086] The distance value sequence is a data sequence composed of the distance values of each color value and the sample color value parameter.

[0087] The Euclidean distance obtained here is used to represent the color difference between the sample and each color system in the target color system subset. It should be noted that by sorting the distance value sequence obtained by traversing the sample color value and the target color system subset from small to large, the first few distance values corresponding to the dyeing formula can be selected as the recommended formula output. The output formula data includes the category and concentration value of the dye.

[0088] See Figure 2 The embodiment also provides a color matching recommendation system, comprising: a database construction module configured to establish a color matching formula history database; a color matching classification module configured to obtain a plurality of color matching data sets based on the color matching formula history database and a preset decision tree classification model; a classification model training module configured to train a pre-constructed LightGBM initial classification model based on the plurality of color matching data sets to obtain a LightGBM classification model; a sample parameter acquisition module configured to obtain sample color value parameters, sample material parameters and sample process parameters according to a sample to be color matched; a target formula screening module configured to obtain a target formula database subset from the color matching formula history database based on the sample material parameters and the sample process parameters; a sample color system acquisition module configured to input the sample color value parameters into the LightGBM classification model, so that the LightGBM classification model outputs a sample color system corresponding to the sample to be color matched; a target color system acquisition module configured to obtain a target color system subset from the target formula database subset based on the sample color system; and a color matching recommendation module configured to obtain a color matching scheme meeting a preset requirement according to the sample color value parameters and the target color system subset.

[0089] The system provided by the embodiment is simple to construct, and ingeniously uses existing formula history data to construct a lightGBM classification model and recommend a formula for a sample to be color matched. The system efficiently groups and screens the formula data by using a decision tree classification model and a LightGBM classification model, and integrates the color matching recommendation method on a cloud platform system for use by a color matching operator, so that the accuracy and efficiency of color matching formula recommendation are improved, and the stability and consistency of dyeing effects are ensured.

[0090] Further, the color matching classification module is configured to obtain a plurality of color matching data sets based on the color matching formula historical database and a preset decision tree classification model, including: extracting material characteristics and process characteristics of different color matching formulas from the color matching formula historical database to obtain a material characteristic set and a process characteristic set; inputting the material characteristic set and the process characteristic set into the decision tree classification model, so that the decision tree classification model splits the color matching formula historical database into a plurality of color matching data sets based on the material characteristic set and the process characteristic set; and the color matching data set is a data set composed of color matching formula historical data with the same material characteristics and the same process characteristics.

[0091] Further, the classification model training module is configured to train a pre-constructed LightGBM initial classification model based on the plurality of color matching data sets to obtain a LightGBM classification model, including: performing data preprocessing based on the color matching formula historical database, and dividing the data set into a training set and a test set; training the pre-constructed LightGBM initial classification model based on the training set and specified parameters to obtain a training model; performing performance evaluation and parameter optimization on the training model based on a preset cross-validation method to obtain a verification model; and evaluating the verification model based on the test set to determine whether the verification model meets preset training requirements, and obtaining the LightGBM classification model if the verification model meets the preset training requirements.

[0092] Further, the color matching recommendation module is configured to obtain a color matching scheme meeting preset requirements based on the sample color value parameter and the target color system subset, including: calculating the Euclidean distance between the sample color value parameter and the color value of each color system in the target color system subset to obtain a distance value sequence; and obtaining a target distance value meeting preset requirements in the distance value sequence, and taking the target color system subset corresponding to the target distance value as the color matching scheme.

[0093] The embodiment is a specific implementation of the color matching recommendation method on a color matching recommendation system. Through the operation of the color matching classification, classification model training, and color matching recommendation modules, an operator can accurately and quickly obtain a color matching recommendation scheme with the help of existing color matching formula historical data, so as to intelligently and automatically perform color matching, improve data management efficiency, and reduce dependence on manual experience.

[0094] The above describes preferred embodiments of the present application. It should be noted that those skilled in the art can make several improvements and refinements without departing from the principles of the present application, and these improvements and refinements are also considered within the protection scope of the present application.

Claims

1. A color matching recommendation method characterized by comprising: The method comprises the following steps: establishing a color matching formula history database; obtaining a plurality of color matching data sets based on the color matching formula history database and a preset decision tree classification model; training a pre-constructed LightGBM initial classification model based on the plurality of color matching data sets to obtain a LightGBM classification model; obtaining sample color value parameters, sample material parameters and sample process parameters according to a sample to be color matched; obtaining a target formula database subset from the color matching formula history database based on the sample material parameters and sample process parameters; inputting the sample color value parameters into the LightGBM classification model to enable the LightGBM classification model to output a sample color system corresponding to the sample to be color matched; obtaining a target color system subset from the target formula database subset based on the sample color system; obtaining a color matching scheme meeting preset requirements according to the sample color value parameters and the target color system subset.

2. The color matching recommendation method according to claim 1, characterized in that, The method comprises the following steps: extracting material characteristics and process characteristics of different color matching formulas from the color matching formula history database to obtain a material characteristic set and a process characteristic set; inputting the material characteristic set and the process characteristic set into the decision tree classification model to enable the decision tree classification model to split the color matching formula history database into a plurality of color matching data sets based on the material characteristic set and the process characteristic set; The color matching data set is a data set composed of color matching formula history data of the same material characteristics and the same process characteristics.

3. The color matching recommendation method according to claim 1, characterized in that, The method comprises the following steps: performing data preprocessing based on the color matching formula history database and dividing the data set into a training set and a test set; training a pre-constructed LightGBM initial classification model based on the training set and specified parameters to obtain a training model; performing performance evaluation and parameter optimization on the training model based on a preset cross-validation method to obtain a validation model; evaluating the validation model based on the test set to determine whether the validation model meets preset training requirements, and obtaining a LightGBM classification model if the validation model meets the preset training requirements.

4. The color matching recommendation method according to claim 3, characterized in that, The method comprises the following steps: If the validation model does not meet the preset training requirements, retraining the validation model based on the training set and specified parameters to obtain a new training model and a new validation model until the new validation model meets the preset training requirements.

5. The color matching recommendation method of claim 1, wherein, The method comprises the following steps: respectively calculating the Euclidean distances between the sample color value parameters and the color value of each color system in the target color system subset to obtain a distance value sequence; obtaining a target distance value meeting preset requirements in the distance value sequence, and taking the target color system subset corresponding to the target distance value as a color matching scheme.

6. The color matching recommendation method according to claim 5, characterized in that, The calculating the Euclidean distance between the sample color value parameter and the color value of each color system in the target color system subset respectively to obtain a distance value sequence comprises: For any color system in the target color system subset, the process of calculating the Euclidean distance between the color value of the color system and the sample color value parameter is: where p is the Euclidean distance, R x , G x , B x are the sample color value parameters, R n , G n , B n are the arbitrary color system color value parameters in the target color system subset; The distance value sequence is a data sequence composed of distance values of each color value and the sample color value parameter.

7. A color matching recommendation system characterized by, A system for performing a color matching recommendation method according to any one of claims 1-6, the system comprising: a database construction module for constructing a color matching formula history database; a color matching classification module for obtaining a plurality of color matching data sets based on the color matching formula history database and a preset decision tree classification model; a classification model training module for training a pre-constructed LightGBM initial classification model based on the plurality of color matching data sets to obtain a LightGBM classification model; a sample parameter acquisition module for acquiring a sample color value parameter, a sample material parameter, and a sample process parameter according to a sample to be color matched; a target formula screening module for obtaining a target formula database subset from the color matching formula history database based on the sample material parameter and the sample process parameter; a sample color system acquisition module for inputting the sample color value parameter into the LightGBM classification model to enable the LightGBM classification model to output a sample color system corresponding to the sample to be color matched; a target color system acquisition module for obtaining a target color system subset from the target formula database subset based on the sample color system; a color matching recommendation module for obtaining a color matching scheme meeting a preset requirement according to the sample color value parameter and the target color system subset.

8. The color matching recommendation system of claim 7, wherein, The color matching classification module is configured to obtain a plurality of color matching data sets based on the color matching formula history database and a preset decision tree classification model, and comprises: extracting material features and process features of different color matching formulas from the color matching formula history database to obtain a material feature set and a process feature set; inputting the material feature set and the process feature set into the decision tree classification model to enable the decision tree classification model to split the color matching formula history database into a plurality of color matching data sets based on the material feature set and the process feature set; The color matching data set is a data set composed of color matching formula history data of the same material feature and the same process feature.

9. The color matching recommendation system of claim 7, wherein, The classification model training module is configured to train a pre-constructed LightGBM initial classification model based on the plurality of color matching data sets to obtain a LightGBM classification model, and comprises: performing data preprocessing based on the color matching formula history database, and dividing the data set into a training set and a test set; training a pre-constructed LightGBM initial classification model based on the training set and specified parameters to obtain a training model; performing performance evaluation and parameter optimization on the training model based on a preset cross-validation method to obtain a validation model; evaluating the validation model based on the test set to determine whether the validation model meets a preset training requirement, and obtaining the LightGBM classification model if the validation model meets the preset training requirement.

10. The color matching recommendation system of claim 7, wherein, The color matching recommendation module is configured to obtain a color matching scheme meeting preset requirements according to the sample color value parameter and the target color system subset, and the color matching scheme includes: Euclidean distances between the sample color value parameter and color value parameters of each color system in the target color system subset are respectively calculated to obtain a distance value sequence; A target distance value meeting preset requirements is obtained from the distance value sequence, and a target color system subset corresponding to the target distance value is taken as the color matching scheme.

Images