A small sample library-based fabric dyeing formula intelligent recommendation system

Abstract

This invention discloses an intelligent recommendation system for fabric dyeing formulas based on a sample library, relating to the field of fabric dyeing technology. When constructing a multi-dimensional sample library, this invention synchronously binds and collects dyeing samples with sampling environment parameters, and collects production environment parameters of the same dimension in real time and incorporates them into formula matching requirements. Based on the environmental deviation between sampling and production environment parameters, it performs targeted optimization of dyeing formulas in the candidate formula group. This invention can fundamentally solve the inherent differences between the sampling environment and the actual production environment, achieving accurate adaptation of sample data to the production scenario without high-cost production environment modifications. By quantitatively compensating for deviations in environmental parameters such as temperature, humidity, and water quality, it reduces color difference and color fastness fluctuations between samples, improving the first-time dyeing success rate. Simultaneously, it stably adapts to complex working conditions such as real-time fluctuations in the workshop environment and continuous batch production, significantly improving the applicability and production stability of dyeing formulas while ensuring the economic efficiency of large-scale production.

Description

Technical Field

[0001] This invention belongs to the field of fabric dyeing technology, specifically a smart recommendation system for fabric dyeing formulas based on a small sample library. Background Technology

[0002] Most existing intelligent recommendation schemes for fabric dyeing formulas rely on a database of historical dyeing samples. They use a single weighted mixed matching search based on target color, fabric type, and process parameters to directly output the corresponding dyeing formula for mass production.

[0003] The existing solution has the following problems: 1) The sample library data is incomplete, only storing dyeing formulas, fabric properties, color indicators and basic process parameters. It cannot reflect the impact of environmental factors such as temperature, humidity and water quality on the final color. The sample data has no environmental traceability, and there is no benchmark reference for the inherent deviations of large and small environments; 2) The matching logic has inherent shortcomings. It uses a single fixed weight of color, type and process parameters for mixed matching. It is necessary to make a trade-off between production adaptability and color accuracy. Either the emphasis on color will make the formula unable to be implemented in production, or the emphasis on environment will cause the color deviation to exceed the standard. Moreover, it only outputs one set of candidate formulas, which is very easy to miss local optimal solutions and there is no room for subsequent optimization and comparison.

[0004] This invention provides an intelligent recommendation system for fabric dyeing formulas based on a small sample library to solve the above-mentioned technical problems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an intelligent recommendation system for fabric dyeing formula based on a small sample library.

[0006] To achieve the above objectives, a first aspect of the present invention provides an intelligent recommendation system for fabric dyeing formulas based on a small sample library, comprising:

[0007] The formula recommendation module is used to generate a multi-dimensional sample library based on several prepared staining samples; extract the generated formula matching requirements; wherein, the staining samples are bound to sampling environment parameters, and the formula matching requirements include staining requirements and production environment parameters; and,

[0008] The matching rules are established based on environmental and color adaptation priorities. Two candidate formula groups are constructed from a multidimensional sample library based on the formula matching requirements and matching rules. The dyeing formulas in the candidate formula groups are optimized using the matching rules. The optimized dyeing formulas are comprehensively evaluated and the target formula is selected.

[0009] In one possible implementation, a multidimensional sample library is generated based on several prepared staining samples, including:

[0010] Several dyeing samples were prepared based on a pre-set sampling standard; environmental parameters during the preparation of the dyeing samples were collected and standardized to serve as sampling environmental parameters.

[0011] Sample data is generated based on staining samples and sampling environment parameters; sample data from several staining samples are integrated into a multidimensional sample library.

[0012] In one possible implementation, sample data is generated based on the staining sample and the sampling environment parameters, including:

[0013] Color features were extracted from stained samples; these features included CIE Lab color data, full-spectrum curves, and color fastness indices.

[0014] The preparation characteristics, color characteristics, and sampling environment parameters of the dyeing sample are integrated into the sample data of the dyeing sample; among which, the preparation characteristics include fabric specifications, dyeing formula, and process parameters.

[0015] In one possible implementation, the generation of recipe matching requirements includes:

[0016] Enter the fabric specifications and dyeing requirements of the fabric to be dyed; collect the environmental parameters of the production workshop in real time, and use them as production environment parameters after standardization; the production environment parameters and the sampling environment parameters are consistent in dimensions.

[0017] Integrate fabric specifications, dyeing requirements, and production environment parameters into formulation matching requirements.

[0018] In one possible implementation, two candidate recipe sets are constructed from a multidimensional sample library based on recipe matching requirements and matching rules, including:

[0019] The matching rules are constructed based on prioritizing environmental and color adaptation.

[0020] Based on the matching rules, matching was performed in the multidimensional sample library to obtain two candidate formulation groups; each candidate formulation group included sample data of several dye samples.

[0021] In one possible implementation, matching rules are constructed with environmental adaptation as the priority. The feature priority in the matching rules is: fabric specification similarity > environmental similarity > color similarity > process similarity, or environmental similarity > color similarity > process similarity.

[0022] In one possible implementation, matching rules are constructed with color adaptation as the priority. The feature priority in the matching rules is: fabric specification similarity > color similarity > process similarity > environment similarity, or color similarity > process similarity > environment similarity.

[0023] In one possible implementation, the staining formulations in the candidate formulation group are optimized using matching rules, including:

[0024] The optimized input data for each sample in the candidate formulation group is constructed based on the matching rules; wherein, the optimized input data includes at least the staining formulation and the environmental bias term.

[0025] The pre-trained formula optimization model is invoked, and the optimized input data is input into the formula optimization model after standardization, and the optimized dyeing formula is output; the formula optimization model is built based on an artificial intelligence model.

[0026] In one possible implementation, optimized input data corresponding to each sample data in the candidate formulation group is constructed based on matching rules, including:

[0027] Determine whether the matching rules include fabric specification similarity;

[0028] Yes, the input data is integrated and optimized by combining dyeing formulas and environmental deviations;

[0029] If not, the fabric specifications of the fabric to be dyed and the sample fabric are used as the specification matching item, and the specification matching item, dyeing formula and environmental deviation item are integrated and optimized into the input data.

[0030] In one possible implementation, the optimized staining formulation is comprehensively evaluated to screen for the target formulation, including:

[0031] The production and dyeing process of the optimized dyeing formula was simulated to obtain simulated dyeing results; among them, the simulated dyeing results include dyeing cost, simulated color difference, color fastness and stability judgment results;

[0032] The data items in the simulated staining results are weighted and summed to obtain a comprehensive score. The staining formula with the highest comprehensive score is taken as the target formula.

[0033] Compared with the prior art, the beneficial effects of the present invention are:

[0034] 1. This invention synchronously binds and collects dyeing samples and sampling environment parameters when constructing a multi-dimensional sample library, and collects production environment parameters of the same dimension in real time and incorporates them into the formula matching requirements; based on the environmental deviation items of the sampling environment and production environment parameters, it performs targeted optimization of dyeing formulas in the candidate formula group. This invention can fundamentally solve the inherent differences between the sampling environment and the actual production environment, and can achieve accurate adaptation of sample data to the production scenario without high-cost modification of the production environment; by quantitatively compensating for deviations in environmental parameters such as temperature, humidity, and water quality, it significantly reduces the color difference and color fastness fluctuations between samples, and improves the first-time dyeing success rate; at the same time, it stably adapts to complex working conditions such as real-time fluctuations in the workshop environment and continuous batch production, and significantly improves the feasibility and production stability of dyeing formulas while ensuring the economics of large-scale production.

[0035] 2. This invention establishes two independent matching rules: one prioritizing environmental adaptability and the other prioritizing color adaptability. It uses a dual-path approach to parallelly search a multi-dimensional sample library, generating two sets of differentiated candidate formulations. After optimizing each set of dyeing formulations, a comprehensive evaluation is conducted to select the target formulation. This invention overcomes the limitations of traditional single-weight matching, which requires trade-offs between production adaptability and color accuracy. It retains two sets of candidate solutions with high feasibility and high color accuracy, avoiding the omission of the optimal solution. The two independent formulations improve production tolerance, preventing production stoppages due to the failure of a single solution. Flexible priority settings adapt to various scenarios such as mass production, high-precision dyeing, and different sample library sizes, resulting in stronger compatibility. The optimized and comprehensively scored formulations balance feasibility, color accuracy, and cost-effectiveness, significantly reducing sampling rework and R&D costs, and comprehensively improving formulation recommendation accuracy, production efficiency, and system applicability. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of the intelligent recommendation process for fabric dyeing formulas in an embodiment of the present invention;

[0038] Figure 2 This is a schematic diagram illustrating the matching and optimization process of dyeing formulations in an embodiment of the present invention. Detailed Implementation

[0039] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] The sampling environment and the dyeing production environment differ significantly. Laboratory sampling typically involves small-scale, precision equipment, small batches, controlled constant temperature and humidity, pure water quality, and standardized, meticulous operations. In contrast, large-scale production involves large industrial dyeing vats, mass production, open workshop environments, recycled water, and continuous batch operations. These differences are multidimensional: First, there are differences in equipment scale. Small-scale sampling requires precise machine-to-bath ratios and high temperature and humidity uniformity, while large dyeing vats have larger volumes, poorer temperature fields and mixing uniformity, resulting in completely different dye adsorption and fixation effects. Second, there are differences in environmental conditions. Laboratories use constant pure water and maintain constant temperature and humidity, while workshops experience significant fluctuations in water hardness and temperature and humidity changes with the seasons and production schedules, directly impacting dye solubility and dyeing rates. Third, there are differences in batch size and operation. Small-scale sampling involves single, meticulous sampling, while large-scale production involves continuous batch processing and dyeing, leading to significant deviations in processing uniformity and process execution. Fourth, there are differences in post-processing. Small-scale sampling involves precise and controllable drying and setting conditions, while large-scale production involves continuous post-processing with low heat uniformity.

[0041] The aforementioned differences are inherent characteristics of large-scale production in the industry and cannot be completely matched by adjusting the production environment to a sampling environment: on the one hand, transforming a large production workshop into a laboratory standard would result in extremely high equipment modification and operating costs, completely losing the economic viability of large-scale production and lacking industrial feasibility; on the other hand, the continuous and large-scale process characteristics of large-scale production are fundamentally contradictory to the characteristics of small-sample, small-batch, and refined production. Even with partial adjustments, the core deviations caused by scale, uniformity, and batch size cannot be eliminated, and color differences between large and small samples will still occur.

[0042] To address the aforementioned issues, this invention provides the following solution: When establishing a multidimensional sample library, sampling environment parameters are simultaneously bound to construct a multidimensional sample library encompassing environment, color, formulation, and process; dual-path independent parallel matching is employed to generate two sets of differentiated candidate formulations prioritizing environment and color; the two sets of candidate formulations are then optimized in conjunction with the actual production environment, and a comprehensive comparison is performed on the optimized formulations to select the target formulation.

[0043] Please see Figure 1The first aspect of this invention provides an intelligent recommendation system for fabric dyeing formulas based on a sample library, including a formula recommendation module: for generating a multi-dimensional sample library based on several prepared dyeing samples; extracting the generated formula matching requirements; wherein the dyeing samples are bound to sampling environment parameters, and the formula matching requirements include dyeing requirements and production environment parameters; and for matching and constructing two candidate formula groups according to matching rules established based on environment adaptation priority and color adaptation priority; optimizing the dyeing formulas in each candidate formula group using the matching rules; and comprehensively evaluating the optimized dyeing formulas to screen and obtain the target formula.

[0044] In addition to the formula recommendation module, the system also includes a database or data acquisition module to collect the data required by the formula recommendation module to make dyeing formula recommendations, and a smart terminal to send the target formula to the staff or the dyeing control module.

[0045] As mentioned above, since the differences between the prototyping environment and the production environment cannot be eliminated, it is necessary to bind the dyed sample to the prototyping environment during prototyping to provide a data foundation for subsequent dual-path parallel matching and targeted environment adaptation optimization.

[0046] Before preparing dyeing samples, a sampling benchmark is set according to the general standards of the textile industry. Several dyeing samples are prepared according to the sampling benchmark, and the sampling environment parameters during the preparation of the dyeing samples are recorded. A sample data can be generated based on the dyeing samples and the sampling environment parameters. The sample data corresponding to several dyeing samples can be integrated to obtain a multidimensional sample library.

[0047] In one example, the process of generating a multidimensional sample library based on several stained samples is as follows:

[0048] X01: Multiple sampling standards are pre-set, and dyeing samples of uniform specifications are prepared based on each sampling standard;

[0049] The sampling benchmark is based on general standards in the textile industry, using fabric specifications, dyeing formulas, and process parameters as variables to prepare dyeing samples of uniform specifications. This means that at least one of the fabric specifications, dyeing formulas, process parameters, or sampling environment parameters differs between each dyeing sample. This process standardizes the preparation process of dyeing samples, ensuring their consistency and repeatability.

[0050] X02: During the preparation of dyeing samples, high-precision sensors are used to collect environmental parameters throughout the entire process of dyeing sample preparation, which are then standardized and used as sampling environmental parameters.

[0051] The sampling environment parameters include temperature, humidity, and air pressure; water hardness / pH / conductivity; and drying / baking parameters. All environmental parameters are standardized before being used as the sampling environment parameters. High-precision sensors are compatible with the data items in the sampling environment parameters. Standardization processing includes normalization and unique thermal encoding to eliminate the influence of different dimensions.

[0052] X03: Extract color features from the prepared staining samples, combine them with the preparation features in the sampling benchmark and the sampling environment parameters to generate a sample data, and integrate the sample data of each staining sample into a multidimensional sample library.

[0053] Color characteristics include CIE Lab color data, full-spectrum curves, and color fastness indices, while preparation characteristics include fabric specifications, dyeing formulations, and process parameters. This multidimensional sample library contains sample data for various dyeing samples, providing a data foundation for matching subsequent dyeing formulation requirements.

[0054] When performing intelligent matching of fabric dyeing formulations, it is necessary to clearly define the formulation matching requirements. The formulation matching requirements should not only include the fabric specifications and color characteristics of the fabric to be dyed, but also determine the production environment parameters used for dyeing the fabric, in order to match a usable and accurate dyeing formulation.

[0055] In one example, the process for generating recipe matching requirements includes:

[0056] M01: Enter the fabric specifications and color requirements of the fabric to be dyed;

[0057] Fabric specifications include fiber composition, the proportion of each fiber, and its coverage. Color requirements include color card number, Lab value, etc. Color requirements can be used to extract CIE Lab color data, full-spectrum curves, and other color parameters. In addition to direct input, color requirements can also be automatically identified and extracted from uploaded color samples.

[0058] M02: Connects to industrial sensors in the production workshop, automatically collects environmental parameters of the production workshop through industrial sensors, and obtains production environment parameters after standardization processing.

[0059] The production environment parameters and the sampling environment parameters are consistent in dimensions, including the temperature, humidity and air pressure of the production workshop, water hardness / pH / conductivity, drying / baking parameters, etc.

[0060] M03: Integrate the fabric specifications, color requirements, and production environment parameters of the fabric to be dyed into a formula matching requirement.

[0061] It is worth noting that the production environment parameters in the formula matching requirements are not the workshop environment parameters when the fabric to be dyed is actually dyed according to the target formula. They are mainly used to match the target formula. If the matched target formula is basically consistent with the production environment parameters in the formula matching requirements, then there is no need to make too many adjustments to the production workshop environment, which helps to reduce dyeing costs.

[0062] Please see Figure 2 After determining the formula matching requirements, several sample data points are matched from the multidimensional sample library based on two priorities: environmental adaptation and color adaptation, respectively, to construct two candidate formula groups. Environmental adaptation priority prioritizes sample data with parameters similar to the production environment during matching in the multidimensional sample library, thus reducing adjustments to the production workshop and lowering costs. Color adaptation priority prioritizes sample data consistent with color requirements during matching in the multidimensional sample library, ensuring dyeing effects and reducing the need for subsequent optimization of the dyeing formula.

[0063] In one example, the process of constructing a candidate formulation group is as follows:

[0064] Z01: Set a matching rule one that prioritizes environmental adaptation. Based on this matching rule one, several candidate samples are matched in the multidimensional sample library. The staining formulas and sampling environment parameters of the candidate samples are integrated into candidate group one.

[0065] Matching rule one is guided by the stability of the production environment, and its matching priority is: environmental similarity > color similarity > process similarity. When matching according to matching rule one, the environmental similarity of the production environment parameters and the sampling environment parameters of all sample data in the multidimensional sample library is calculated. Sample data with environmental similarity greater than the preset environmental similarity threshold are integrated into candidate pool one. Then, similarity calculation and matching are performed on the sample data in candidate pool one according to color features and process parameters. The matching results are sorted according to priority, and 3-5 candidate samples are selected. The dyeing formula and sampling environment parameters of the corresponding sample data of the candidate samples are integrated into candidate group one.

[0066] Z02: Set a matching rule two that prioritizes color matching. Based on this matching rule two, several dyeing formulas are matched in the multidimensional sample library. These dyeing formulas are then integrated into candidate group two.

[0067] Matching rule two is color accuracy-oriented, with the following matching priority: color similarity > process similarity > environmental similarity. When matching according to matching rule two, color similarity is calculated between the color features and the color features of all sample data in the multidimensional sample library. Sample data with color similarity greater than a preset color similarity threshold are integrated into candidate pool two. Then, matching is performed in candidate pool two according to process parameters and production environment parameters. The final matching results are sorted by priority, and 3-5 candidate samples are selected. The dyeing formula and sampling environment parameters in the corresponding sample data of these candidate samples are integrated into candidate group two.

[0068] Z03: Perform slight deduplication on the repeated staining formulas in candidate group 1 and candidate group 2 to obtain two formula candidate groups.

[0069] Mild deduplication refers to removing duplicate dyeing formulas from candidate group one or candidate group two. During deletion, priority should be given to retaining both candidate groups to avoid leaving any candidate group empty of dyeing formulas after mild deduplication. This process strictly preserves the independence and differences between the two candidate sets, avoiding subsequent redundant optimization, while ensuring that each candidate set covers the two major advantages of ease of production implementation and high color accuracy.

[0070] It is worth noting that when the above matching is used to obtain the candidate formula group, the fabric to be dyed and the fabric specifications in the sampling benchmark are not considered. This improves the efficiency of matching sample data with priority given to environmental or color adaptation. Relatively speaking, it does not require too much sample data in the multidimensional sample library. Therefore, it is suitable for scenarios with limited data, especially when there are few fabric specifications in the sampling benchmark.

[0071] However, when obtaining candidate formulas without considering fabric specifications, the impact of different fabric specifications on the actual dyeing effect is not taken into account. In order to ensure the final dyeing effect of the fabric to be dyed, the fabric specifications need to be used as one of the optimization bases of the dyeing formula so that the final optimized dyeing formula can be applied to the fabric specifications of the fabric to be dyed.

[0072] Fabric specifications can also be included in the matching rules. Matching rule one is fabric specification similarity > environmental similarity > color similarity > process similarity, and matching rule two is fabric specification similarity > color similarity > process similarity > environmental similarity.

[0073] If the fabric specifications of the fabric to be dyed are included in the matching rules, the matching dyeing formula will have a high compatibility with the fabric. Therefore, the fabric specifications can be disregarded in subsequent dyeing formula optimization, improving the optimization efficiency of the dyeing formula. However, if the fabric specifications are included in the matching rules, it is necessary to ensure that the multidimensional sample library contains a large amount of sample data of different fabric specifications. Therefore, it is suitable for scenarios with a large amount of sample data in the multidimensional sample library.

[0074] After determining two candidate formulation groups, the dyeing formulations were independently optimized mainly based on the deviations between the sampling environment parameters and the production environment parameters. Production simulation was performed on the optimized dyeing formulations, and the optimized dyeing formulations were comprehensively scored based on the production simulation results. The dyeing formulation with the highest comprehensive score was selected as the target formulation.

[0075] In one example, the process of determining the target formulation is as follows:

[0076] M01: For each dyeing formula in the two candidate formula groups, calculate the deviation value of each data item in the sampling environment parameter and the production environment parameter, and use it as the environmental deviation item; if necessary, associate the fabric specifications of the fabric to be dyed with the fabric specifications of the corresponding sampling fabric of each dyeing formula to establish a specification matching item.

[0077] M02: Call the pre-trained formula optimization model; input the environmental deviation term and dyeing formula (including specification matching term if necessary) into the formula optimization model after standardization, and output the optimized dyeing formula;

[0078] M03: The dyeing process of the dyeing formula is simulated under production environment parameters using software to obtain simulated dyeing results; the simulated dyeing results include dyeing cost, simulated color difference, color fastness and stability prediction results;

[0079] M04: Calculate the overall score of the remaining optimized dyeing formulas and select the dyeing formula with the highest overall score as the target formula.

[0080] When calculating the deviation values ​​of each data item in the sampling environment parameters and the production environment parameters, if a certain data item in the production environment parameters can be adjusted to the corresponding sampling environment parameters at low cost, then the environmental deviation item for that data item does not need to be calculated. During subsequent actual dyeing, this data item can be adjusted to be consistent with the sampling environment parameters. For example, drying / baking parameters, equipment speed, etc., can have their corresponding environmental deviation items set to 0.

[0081] If a certain data point in the production environment cannot be adjusted, or cannot be adjusted at low cost to match the corresponding sampling environment parameters, then it is necessary to calculate the environmental deviation item for that data point and optimize the dyeing formula based on these environmental deviation items. For example, parameters such as temperature, humidity, air pressure, water hardness / pH / conductivity are difficult to adjust to match the sampling environment parameters, so it is necessary to calculate the environmental deviation items for these parameters.

[0082] As mentioned above, specification matching items only need to be set when necessary. That is, if the fabric specifications are not included in the matching rules, then specification matching items need to be set; otherwise, only the environmental deviation item needs to be calculated.

[0083] The formulation optimization model can be based on mature artificial intelligence models such as deep convolutional neural networks and backpropagation neural networks. It can adopt existing mature model structures and only needs to be trained by aligning a large amount of data. The training dataset includes input data and corresponding output data. The dimensions of the input data and the optimization input data are consistent, and the dimensions of the output data are consistent with the dimensions of the staining formulation.

[0084] During AI model training, the input data should have pre-defined specification matching terms. If no specification matching terms are available during actual optimization, these terms should be set to 0 in the optimized input data. Other terms similar to specification matching terms should be processed according to the same logic. The training process and parameter settings can be found in existing solutions and will not be elaborated here.

[0085] It is worth noting that when simulating the dyeing process under production environment parameters, these parameters may differ from those specified in the formula matching requirements. For example, if the actual production drying temperature is A℃, and the sampling environment parameters specify a drying temperature of B℃, and the drying equipment can be adjusted to B℃, then A℃ in the production environment parameters needs to be replaced with B℃.

[0086] The optimization of dyeing formulations mainly includes dye concentration, chelating agent dosage, and temperature rise curves, and may also include other optimizable items. The software modeling process can utilize Datacolor Match Textile software to perform virtual dyeing simulations of the optimized dyeing formulation.

[0087] After obtaining the simulated dyeing results, the data items in the simulated dyeing results are weighted and fused using pre-set weighting coefficients to obtain a comprehensive score for each dyeing formula. The dyeing formula with the highest comprehensive score is selected as the target formula. Next, actual production is carried out on the fabric to be dyed based on the target formula. In actual production, it is necessary to adjust the actual production parameters according to the sampling environment parameters to ensure that the actual dyeing effect meets expectations.

[0088] In one example, the simulated dyeing results include dyeing cost, simulated color difference, color fastness, and stability assessment results, with preset weighting coefficients. Specifically, the values ​​were set to 0.3, 0.3, 0.2, and 0.2 respectively, and the data items in the simulated staining results were normalized and marked as follows. , Number the data items. Indicates the cost of dyeing. Indicates simulated color difference. Indicates color fastness, Indicate the stability determination result; using the formula .

[0089] It should be noted that the weight coefficients of the data items in the above examples need to be set in combination with experience and coloring requirements, and the values ​​of the weight coefficients are not fixed here.

[0090] It is worth noting that if the actual dyeing effect of the target formula differs from the color difference of the dyed sample by more than a preset threshold, the formula optimization model can be updated using this set of data to improve the optimization accuracy of the formula optimization model.

[0091] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments.

[0092] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any other combination thereof. When implemented using a software program, it can be implemented entirely or partially in the form of a computer program product. This computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device containing one or more servers, data centers, etc., that can be integrated with the medium. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state disks (SSDs)).

[0093] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A smart recommendation system for fabric dyeing formulas based on a small sample library, characterized in that, include: Formula recommendation module: used to generate a multidimensional sample library based on several prepared staining samples; Extract the generated formula matching requirements; among them, the dyeing sample is bound to sampling environment parameters, and the formula matching requirements include dyeing requirements and production environment parameters; and, The matching rules are established based on environmental and color adaptation priorities. Two candidate formula groups are constructed from a multidimensional sample library based on the formula matching requirements and matching rules. The dyeing formulas in the candidate formula groups are optimized using the matching rules. The optimized dyeing formulas are comprehensively evaluated and the target formula is selected.

2. The intelligent fabric dyeing recipe recommendation system based on small sample library according to claim 1, characterized in that, A multidimensional sample library was generated based on several prepared staining samples, including: Several dyeing samples are prepared based on a pre-set sampling standard; during the preparation of the several dyeing samples, environmental parameters during the preparation process are collected and used as sampling environmental parameters after standardization. Sample data is generated based on the staining samples and the sampling environment parameters; the sample data of several staining samples are integrated into a multidimensional sample library.

3. The intelligent recommendation system for fabric dyeing formulas based on a small sample library according to claim 2, characterized in that, Sample data is generated based on the staining samples and the sampling environment parameters, including: Color features are extracted from the stained samples; these color features include CIE Lab color data, full-spectrum curves, and color fastness indices. The preparation characteristics, color characteristics, and sampling environment parameters of the dyed sample are integrated into the sample data of the dyed sample; wherein, the preparation characteristics include fabric specifications, dyeing formula, and process parameters.

4. The intelligent fabric dyeing recipe recommendation system based on small sample library according to claim 1, characterized in that, The generation of recipe matching requirements includes: Enter the fabric specifications and dyeing requirements of the fabric to be dyed; collect the environmental parameters of the production workshop in real time, and use them as production environment parameters after standardization; the production environment parameters and the sampling environment parameters are consistent in dimensions. The fabric specifications, dyeing requirements, and production environment parameters are integrated into the formula matching requirements.

5. The intelligent recommendation system for fabric dyeing formulas based on a small sample library according to claim 1, characterized in that, Based on the recipe matching requirements and matching rules, two candidate recipe groups are constructed from a multidimensional sample library, including: The matching rules are constructed based on prioritizing environmental and color adaptation. Based on the matching rules, matching is performed in the multidimensional sample library to obtain two candidate formulation groups; wherein each candidate formulation group includes sample data of several dye samples.

6. A smart recommendation system for fabric dyeing formulas based on a small sample library as described in claim 1 or 5, characterized in that, Matching rules are constructed with environmental adaptability as the priority. The feature priority in the matching rules is: fabric specification similarity > environmental similarity > color similarity > process similarity, or environmental similarity > color similarity > process similarity.

7. A smart recommendation system for fabric dyeing formulas based on a small sample library as described in claim 1 or 5, characterized in that, Matching rules are constructed with color matching as the priority. The feature priority in the matching rules is: fabric specification similarity > color similarity > process similarity > environment similarity, or color similarity > process similarity > environment similarity.

8. The intelligent fabric dyeing recipe recommendation system based on small sample library according to claim 1, characterized in that, Optimize staining formulations in the candidate formulation group using matching rules, including: The optimized input data for each sample in the candidate formulation group is constructed based on the matching rules; wherein, the optimized input data includes at least the staining formulation and the environmental bias term. The pre-trained formula optimization model is invoked, and the optimized input data is input into the formula optimization model after standardization, and the optimized dyeing formula is output; wherein, the formula optimization model is constructed based on an artificial intelligence model.

9. The intelligent recommendation system for fabric dyeing formulas based on a small sample library according to claim 8, characterized in that, Based on matching rules, optimized input data corresponding to each sample data in the candidate formulation group is constructed, including: Determine whether the matching rules include fabric specification similarity; Yes, the input data is integrated and optimized by combining dyeing formulas and environmental deviations; If not, the fabric specifications of the fabric to be dyed and the sample fabric are used as the specification matching item, and the specification matching item, dyeing formula and environmental deviation item are integrated and optimized into the input data.

10. The intelligent recommendation system for fabric dyeing formulas based on a small sample library according to claim 1, characterized in that, The optimized dyeing formulas were comprehensively evaluated, and target formulas were selected, including: The production and dyeing process of the optimized dyeing formula was simulated to obtain simulated dyeing results; among them, the simulated dyeing results include dyeing cost, simulated color difference, color fastness and stability judgment results; The data items in the simulated staining results are weighted and summed to obtain a comprehensive score, and the staining formula with the highest comprehensive score is taken as the target formula.

Images