An intelligent classification system for simulated material shapes based on machine learning

Abstract

The present invention discloses an intelligent classification system for simulated material shapes based on machine learning. The system comprises: a data acquisition module that extracts the simulated material data set required for a user's design and can export it to generate corresponding image data; an intelligent classification algorithm module that selects a neural network model algorithm suitable for the MCD system by improving the current neural network algorithm based on the MCD simulated material data set; and a user interface module that intelligently classifies the UI interface module on the system interface, allowing users to interact with data to ensure that the designed module structure can be executed according to the user's logic. The present invention transforms the acquisition of object data structure information into the acquisition of object information through image recognition, and demonstrates the feasibility of intelligent classification of simulated material shapes based on machine learning through simulation experiments. The design can be applied to MCD systems.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent classification, and in particular to an intelligent classification system for simulated material shapes based on machine learning. Background Art

[0002] With the development of artificial intelligence and machine learning technologies, intelligent classification systems based on machine learning have been widely used in various fields. Although existing technologies have made some progress in the intelligent classification of simulated material shapes, there are still some shortcomings, including the following deficiencies:

[0003] 1. Existing technologies in intelligent classification algorithm modules may be limited by traditional classification algorithm models and have difficulty adapting to the complex multi-dimensional data characteristics of MCD systems. In contrast, our technical solution uses a machine learning-based algorithm, which can more flexibly select a neural network model algorithm suitable for the MCD system, thereby improving classification accuracy and generalization capabilities;

[0004] 2. In traditional systems, simulation functions may be limited by design and simulation mode units. Users cannot flexibly adjust the logical structure of modeling, making it difficult to meet complex simulation requirements.

[0005] To this end, we propose an intelligent classification system for simulated material shapes based on machine learning to solve the above problems. Summary of the Invention

[0006] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0007] In view of the above-mentioned problems existing in the existing intelligent classification system of simulated material shapes based on machine learning, the present invention is proposed.

[0008] Therefore, the purpose of the present invention is to provide an intelligent classification system for simulated material shapes based on machine learning, which converts the acquisition of object data structure information into information about the object through image recognition, and demonstrates the feasibility of intelligent classification of simulated material shapes based on machine learning through simulation experiments.

[0009] To solve the above technical problems, the present invention provides the following technical solutions: an intelligent classification system for simulated material shapes based on machine learning, comprising:

[0010] Data acquisition module: extracts the simulation material data set required by the user's design and can export it to generate corresponding image data;

[0011] Intelligent classification algorithm module: This module is based on the data set of MCD simulation materials and selects a neural network model algorithm suitable for the MCD system by improving the current neural network algorithm;

[0012] User interface module: This module is an intelligently classified UI interface module on the system interface. Users interact with data on this module to ensure that the designed module structure can be executed according to the user's logic;

[0013] Simulation function module: includes design mode unit and simulation mode unit. The design mode unit is the interface module for design users. Users continuously adjust the logical structure of the model in the design mode unit. The system will save the user-designed logical structure in turn to ensure that the data can be persisted. The simulation mode unit implements the user-designed logical structure through dynamic simulation.

[0014] File storage module: This module is used to store corresponding data parameters, including the data set obtained by the data acquisition module, or some default parameter information.

[0015] As a preferred solution of the intelligent classification system of simulated material shapes based on machine learning of the present invention, the design pattern unit includes four submodules:

[0016] Functional interface submodule, which manages the various components displayed on the user interface. Its purpose is to obtain relevant information set by the user and temporarily cache it in the system;

[0017] Data acquisition submodule: Since the data set of the simulation material designed by the user is in the MCD system, this module is used to quickly obtain the data set of the simulation material;

[0018] The picture interception submodule is used to intercept the data of a single picture and is a subordinate module of the data acquisition submodule;

[0019] The data storage submodule can save the data structure transmitted from the user interface. When the function interface submodule exits, the system will create a corresponding data storage submodule to store the data for a long time.

[0020] As a preferred solution of the intelligent classification system for simulated material shapes based on machine learning of the present invention, the simulation mode unit includes the following submodules:

[0021] Data real-time acquisition module, which is used to obtain the image information required for classification in the simulation process in real time;

[0022] Trigger the image sensor module. During the simulation process, the module will continuously detect whether it is triggered. When the module is triggered, it will call the image capture submodule to obtain the image information of the current simulation material and cache it;

[0023] Simulation process classification module, which consists of three small sub-modules;

[0024] Data recording module, which is used to record various information generated during the simulation process. The data will be used for system testing of the intelligent classification function module of simulated material shape.

[0025] As a preferred solution of the intelligent classification system for simulated material shapes based on machine learning of the present invention, the simulation process classification module includes the following:

[0026] Neural network reading module, which reads the neural network model in the system;

[0027] Execute the classification module. During the simulation process, the module will continuously detect whether it is triggered. When the module is triggered, it will obtain the cached image information from the system, then call the neural network algorithm, pass the image information into the neural network model, obtain the corresponding classification label, and achieve the classification effect;

[0028] The matrix conversion module converts the position of the simulated materials in the system. The effect on the display interface is to transfer the corresponding simulated materials to the classified port positions.

[0029] As a preferred embodiment of the intelligent classification system for simulated material shapes based on machine learning of the present invention, the method for intelligent classification of simulated material shapes based on machine learning includes the following steps:

[0030] Step 1: Data collection and preparation: Collect a set of multimodal data containing different simulated material shapes; label the data and assign the correct category label to each sample;

[0031] Step 2: Feature extraction and selection: Extract features from the data of each modality and select the most representative features based on their importance and relevance.

[0032] Step 3: Data fusion: Use multimodal fusion technology to fuse data from different modalities to obtain more comprehensive information;

[0033] Step 4: Model selection and training: Select a convolutional neural network, recurrent neural network, or other traditional machine learning model, divide the data into a training set and a test set, train the model with the training set, evaluate the model's performance with the test set, and use an optimization algorithm to train the model;

[0034] Step 5: Model evaluation and optimization: Evaluate the trained model and optimize the model based on the evaluation results;

[0035] Step 6: Edge computing optimization: Deploy the trained model to edge devices for efficient operation;

[0036] Step 7: Dynamic simulation test: Integrate the intelligent classification model in the MCD system and provide a user interface to allow users to upload simulated material data for classification. Based on the data uploaded by the user, call the model for classification prediction and display the prediction results.

[0037] As a preferred solution of the intelligent classification system for simulated material shapes based on machine learning of the present invention, the algorithm formula used in step 4 is:

[0038]

[0039] Where N represents the number of samples in the data set; D i represents the data of the i-th sample; Ω represents the spatial range of the data sample; f(D i ) represents the sample D i Perform feature extraction; α i represents the feature weight parameter, corresponding to the feature of the i-th sample; g is the classifier function, which normalizes the feature vector and adds the classifier parameters for classification; β is the normalization parameter, which is used to adjust the degree of normalization; γ is the classifier parameter, which is used to adjust the behavior of the classifier; c is the classification result.

[0040] As a preferred solution of the intelligent classification system for simulated material shapes based on machine learning of the present invention, wherein: the classification result is obtained by using the formula value range in step 5, and the specific steps are as follows:

[0041] S1. Prepare the dataset: First prepare the dataset D;

[0042] S2, feature extraction: for each sample D in the dataset i Perform feature extraction and obtain the feature vector f(D i );

[0043] S3, normalization processing: normalizing the eigenvector to obtain a normalized eigenvector;

[0044] S4. Calculate the range of the formula: Substitute the normalized eigenvector into the formula and calculate the range of the formula:

[0045]

[0046] S5. Determine the classification result: Determine the classification result based on the formula value range, select the category that makes the formula value range the largest as the classification result, and find the category c that makes g(…) the largest.

[0047] As a preferred solution of the intelligent classification system of simulated material shapes based on machine learning of the present invention, it is assumed that a circular data sample is input in step S5 and the formula is calculated:

[0048] g(…)=0.8

[0049] If the threshold is 0.5, the sample is classified as a circle because g(...)=0.8>0.5.

[0050] As a preferred solution of the intelligent classification system of simulated material shapes based on machine learning described in the present invention, the performance test in the MCD system in step seven is divided into two parts of testing, one is the performance test of the functional interface module under the design pattern unit, and the other is the test triggering image sensor module and simulation process classification module.

[0051] As a preferred solution of the intelligent classification system of simulated material shapes based on machine learning described in the present invention, the test method in step seven is mainly a test at the code level, so the current time is recorded on each functional module, and then the current time is recorded after the functional module ends. By comparing the two time nodes, the time required for the module is obtained, and the performance of the functional module is tested based on this time.

[0052] Beneficial effects of the present invention:

[0053] 1. For simulated material data, this invention transforms the acquisition of object data structure information into obtaining object information through image recognition, and demonstrates the feasibility of intelligent classification of simulated material shapes based on machine learning through simulation experiments;

[0054] 2. From the perspective of data set acquisition, the present invention designs a data acquisition module that can quickly acquire the data set of simulated material shapes created by the user. This design can be applied to the MCD system;

[0055] 3. The algorithm proposed in this invention has significant advantages in the classification of simulated materials. Data shows that the model has high accuracy in identifying simulated materials and outperforms other intelligent classification models. Furthermore, the model has fast training convergence and requires less time and data iterations. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0057] Figure 1 Schematic diagram of the intelligent classification module of the intelligent classification system for simulated material shapes based on machine learning of the present invention. DETAILED DESCRIPTION

[0058] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0059] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0060] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0061] Furthermore, the present invention is described in detail with reference to schematic diagrams. For ease of illustration, when describing the embodiments of the present invention, cross-sectional views illustrating device structures may be partially enlarged and not to scale. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of protection of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0062] Reference Figure 1 , provides an intelligent classification system for simulated material shapes based on machine learning, including:

[0063] Data acquisition module: extracts the simulation material data set required by the user's design and can export it to generate corresponding image data;

[0064] Intelligent classification algorithm module: This module is based on the data set of MCD simulation materials and selects a neural network model algorithm suitable for the MCD system by improving the current neural network algorithm;

[0065] User interface module: This module is an intelligently classified UI interface module on the system interface. Users interact with data on this module to ensure that the designed module structure can be executed according to the user's logic;

[0066] Simulation function module: includes design mode unit and simulation mode unit. The design mode unit is the interface module for design users. Users continuously adjust the logical structure of the model in the design mode unit. The system will save the user-designed logical structure in turn to ensure that the data can be persisted. The simulation mode unit implements the user-designed logical structure through dynamic simulation.

[0067] File storage module: This module is used to store corresponding data parameters, including the data set obtained by the data acquisition module, or some default parameter information.

[0068] The design pattern unit includes four submodules:

[0069] Functional interface submodule, which manages the various components displayed on the user interface. Its purpose is to obtain relevant information set by the user and temporarily cache it in the system;

[0070] Data acquisition submodule: Since the data set of the simulation material designed by the user is in the MCD system, this module is used to quickly obtain the data set of the simulation material;

[0071] The picture interception submodule is used to intercept the data of a single picture and is a subordinate module of the data acquisition submodule;

[0072] The data storage submodule can save the data structure transmitted from the user interface. When the function interface submodule exits, the system will create a corresponding data storage submodule to store the data for a long time.

[0073] The simulation mode unit includes the following submodules:

[0074] Data real-time acquisition module, which is used to obtain the image information required for classification in the simulation process in real time;

[0075] Trigger the image sensor module. During the simulation process, the module will continuously detect whether it is triggered. When the module is triggered, it will call the image capture submodule to obtain the image information of the current simulation material and cache it;

[0076] Simulation process classification module, which consists of three small sub-modules;

[0077] Data recording module, which is used to record various information generated during the simulation process. The data will be used for system testing of the intelligent classification function module of simulated material shape.

[0078] Furthermore, the simulation process classification module includes the following:

[0079] Neural network reading module, which reads the neural network model in the system;

[0080] Execute the classification module. During the simulation process, the module will continuously detect whether it is triggered. When the module is triggered, it will obtain the cached image information from the system, then call the neural network algorithm, pass the image information into the neural network model, obtain the corresponding classification label, and achieve the classification effect;

[0081] The matrix conversion module converts the position of the simulated materials in the system. The effect on the display interface is to transfer the corresponding simulated materials to the classified port positions.

[0082] The method for intelligent classification of simulated material shapes based on machine learning includes the following steps:

[0083] Step 1: Data collection and preparation: Collect a set of multimodal data containing different simulated material shapes; label the data and assign the correct category label to each sample;

[0084] Step 2: Feature extraction and selection: Extract features from the data of each modality and select the most representative features based on their importance and relevance.

[0085] Step 3: Data fusion: Use multimodal fusion technology to fuse data from different modalities to obtain more comprehensive information;

[0086] Step 4: Model selection and training: Select a convolutional neural network, recurrent neural network, or other traditional machine learning model, divide the data into a training set and a test set, train the model with the training set, evaluate the model's performance with the test set, and use an optimization algorithm to train the model;

[0087] Step 5: Model evaluation and optimization: Evaluate the trained model and optimize the model based on the evaluation results;

[0088] Step 6: Edge computing optimization: Deploy the trained model to edge devices for efficient operation;

[0089] Step 7: Dynamic simulation test: Integrate the intelligent classification model in the MCD system and provide a user interface to allow users to upload simulated material data for classification. Based on the data uploaded by the user, call the model for classification prediction and display the prediction results.

[0090] Specifically, the algorithm formula used in step 4 is:

[0091]

[0092] Where N represents the number of samples in the data set; D i represents the data of the i-th sample; Ω represents the spatial range of the data sample; f(D i ) represents the sample D i Perform feature extraction; α i represents the feature weight parameter, corresponding to the feature of the i-th sample; g is the classifier function, which normalizes the feature vector and adds the classifier parameters for classification; β is the normalization parameter, which is used to adjust the degree of normalization; γ is the classifier parameter, which is used to adjust the behavior of the classifier; c is the classification result.

[0093] More specifically, in step 5, the classification result is obtained through the formula value range, and the specific steps are as follows:

[0094] S1. Prepare the dataset: First prepare the dataset D;

[0095] S2, feature extraction: for each sample D in the dataset i Perform feature extraction and obtain the feature vector f(D i );

[0096] S3, normalization processing: normalizing the eigenvector to obtain a normalized eigenvector;

[0097] S4. Calculate the range of the formula: Substitute the normalized eigenvector into the formula and calculate the range of the formula:

[0098]

[0099] S5. Determine the classification result: Determine the classification result based on the formula range, select the category that maximizes the formula range as the classification result, and find the category c that maximizes g(…). Assume that a circular data sample is input in step S5 and the formula is calculated:

[0100] g(…)=0.8

[0101] If the threshold is 0.5, the sample is classified as a circle because g(...)=0.8>0.5.

[0102] Among them, the performance test in the MCD system in step seven is divided into two parts of testing. One is the performance test of the functional interface module under the design pattern unit, and the other is the test triggering the image sensor module and the simulation process classification module. The test method in step seven: This part is mainly tested at the code level, so the current time is recorded on each functional module, and then the current time is recorded after the functional module ends. By comparing the two time nodes, the time required for the module is obtained, and the performance test of the functional module is performed based on this time. The specific test is as follows:

[0103] Test Results: As shown in the table, each sub-function module was repeatedly operated more than 20 times, and the average test results were obtained. It can be seen that it takes approximately 668ms to start the function interface module, which means that a delay of 0.668s is required to open a function interface. This is within the normal and acceptable delay range, and the function interface module performance meets the requirements.

[0104] Table MCD performance test results

[0105]

[0106] The triggering image sensor module experienced a delay of 181.66ms. This delay is due to the need to load the image information for the current display interface. In the previous design, this module was separated from the simulation process classification module, so this delay is within a reasonable range. Finally, the simulation process classification module experienced an average delay of only 40.14ms, which is essentially unnoticeable to users during dynamic simulations. Therefore, this performance test met the requirements.

[0107] This paper applies machine learning technology to the MCD system and proposes a hybrid model algorithm suitable for the material shape classification of the MCD system simulation. Based on this algorithm, an intelligent classification system for industrial intelligent simulation pipelines is designed and implemented. Secondly, through simulation experiments, the following conclusions are drawn:

[0108] 1. For simulated material data, this invention transforms the acquisition of object data structure information into obtaining object information through image recognition, and demonstrates the feasibility of intelligent classification of simulated material shapes based on machine learning through simulation experiments;

[0109] 2. From the perspective of data set acquisition, the present invention designs a data acquisition module that can quickly acquire the data set of simulated material shapes created by the user. This design can be applied to the MCD system;

[0110] 3. The algorithm proposed in the present invention has great advantages in the classification of simulated materials. Data shows that the model has high accuracy in the identification of simulated materials and is superior to other intelligent classification models. At the same time, the model training converges quickly, consumes less time and requires fewer data iterations.

[0111] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent classification system for simulated material shapes based on machine learning, characterized in that: include: Data acquisition module: extracts the simulation material data set required by the user's design and can export it to generate corresponding image data; Intelligent classification algorithm module: This module is based on the data set of MCD simulation materials and selects a neural network model algorithm suitable for the MCD system by improving the current neural network algorithm; User interface module: This module is an intelligently classified UI interface module on the system interface. Users interact with data on this module to ensure that the designed module structure can be executed according to the user's logic; Simulation function module: includes design mode unit and simulation mode unit. The design mode unit is the interface module for design users. Users continuously adjust the logical structure of the model in the design mode unit. The system will save the user-designed logical structure in turn to ensure that the data can be persisted. The simulation mode unit implements the user-designed logical structure through dynamic simulation. File storage module: This module is used to store corresponding data parameters, including the data set obtained by the data acquisition module, or some default parameter information; The method for intelligent classification of simulated material shapes based on machine learning includes the following steps: Step 1: Data collection and preparation: Collect a set of multimodal data containing different simulated material shapes; label the data and assign the correct category label to each sample; Step 2: Feature extraction and selection: Extract features from the data of each modality and select the most representative features based on their importance and relevance. Step 3: Data fusion: Use multimodal fusion technology to fuse data from different modalities to obtain more comprehensive information; Step 4: Model selection and training: Select a convolutional neural network, recurrent neural network, or other traditional machine learning model, divide the data into a training set and a test set, train the model with the training set, evaluate the model's performance with the test set, and use an optimization algorithm to train the model; Step 5: Model evaluation and optimization: Evaluate the trained model and optimize the model based on the evaluation results; Step 6: Edge computing optimization: Deploy the trained model to edge devices for efficient operation; Step 7: Dynamic simulation test: Integrate the intelligent classification model into the MCD system and provide a user interface to allow users to upload simulated material data for classification. Based on the data uploaded by the user, the model is called to perform classification prediction and display the prediction results. The algorithm formula used in step 4 is: in, Indicates the number of samples in the dataset; Represents the data of the i-th sample; Indicates the spatial extent of the data sample; Indicates the sample Perform feature extraction; represents the feature weight parameter, corresponding to the feature of the i-th sample; g is the classifier function, which normalizes the feature vector and adds the classifier parameter for classification; is the normalization parameter, which is used to adjust the degree of normalization; is a classifier parameter used to adjust the behavior of the classifier; is the classification result; In step 5, the classification result is obtained through the formula value range. The specific steps are as follows: S1. Prepare the dataset: First prepare the dataset D; S2, feature extraction: for each sample in the dataset Perform feature extraction to obtain feature vector ; S3, normalization processing: normalizing the eigenvector to obtain a normalized eigenvector; S4. Calculate the range of the formula: Substitute the normalized eigenvector into the formula and calculate the range of the formula: ； S5. Determine the classification result: Determine the classification result based on the formula value range, select the category that makes the formula value range the largest as the classification result, and find the category that makes g The largest category ; Assume that a circular data sample is input in step S5 and is calculated using the formula: g =0.8 If the threshold is 0.5, then the sample is classified as a circle because g =0.8 0.5; The performance test in the MCD system in step seven is divided into two parts: one is the performance test of the functional interface module under the design pattern unit, and the other is the test triggering image sensor module and simulation process classification module; The test method in step seven: This part mainly tests at the code level, so the current time is recorded on each functional module, and then the current time is recorded after the functional module ends. By comparing the two time nodes, the time required for the module is obtained, and the performance test of the functional module is performed based on this time.

2. The intelligent classification system for simulated material shapes based on machine learning according to claim 1, characterized in that: The design pattern unit includes four sub-modules: Functional interface submodule, which manages the various components displayed on the user interface. Its purpose is to obtain relevant information set by the user and temporarily cache it in the system; Data acquisition submodule: Since the data set of the simulation material designed by the user is in the MCD system, this module is used to quickly obtain the data set of the simulation material; The picture interception submodule is used to intercept the data of a single picture and is a subordinate module of the data acquisition submodule; The data storage submodule can save the data structure transmitted from the user interface. When the function interface submodule exits, the system will create a corresponding data storage submodule to store the data for a long time.

3. The intelligent classification system for simulated material shapes based on machine learning according to claim 2, characterized in that: The simulation mode unit includes the following submodules: Data real-time acquisition module, which is used to obtain the image information required for classification in the simulation process in real time; Trigger the image sensor module. During the simulation process, the module will continuously detect whether it is triggered. When the module is triggered, it will call the image capture submodule to obtain the image information of the current simulation material and cache it; Simulation process classification module, which consists of three small sub-modules; Data recording module, which is used to record various information generated during the simulation process. The data will be used for system testing of the intelligent classification function module of simulated material shape.

4. The intelligent classification system for simulated material shapes based on machine learning according to claim 3, characterized in that: The simulation process classification module includes the following: Neural network reading module, which reads the neural network model in the system; Execute the classification module. During the simulation process, the module will continuously detect whether it is triggered. When the module is triggered, it will obtain the cached image information from the system, then call the neural network algorithm, pass the image information into the neural network model, obtain the corresponding classification label, and achieve the classification effect; The matrix conversion module converts the position of the simulated materials in the system. The effect on the display interface is to transfer the corresponding simulated materials to the classified port positions.

Images