Environmental micro-plastic classification method and equipment based on meta-learning, medium and product
By preprocessing and classifying the infrared spectral information of microplastics using a meta-learning-based deep neural network model, the problem of accuracy in classifying microplastics in the environment is solved, enabling rapid and accurate identification and monitoring of microplastics.
Patent Information
- Application Number
- CN202411020839.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-12
AI Technical Summary
Existing technologies struggle to accurately identify and classify various microplastics in the environment, making it impossible to effectively monitor and assess their pollution levels and distribution, and thus hinder the development of effective environmental protection strategies.
A deep neural network model was trained using a meta-learning approach. By acquiring and preprocessing the infrared spectral information of microplastics, an environmental microplastic classification model was constructed. This model was then used to determine the category and confidence level of microplastics.
It enables rapid and accurate classification of environmental microplastics, improves the accuracy and adaptability of classification, and allows for better monitoring and assessment of microplastic pollution.
Smart Images

Figure CN118940091A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of environmental microplastic classification, and in particular to a meta-learning-based environmental microplastic classification method, equipment, medium and product. Background Art
[0002] Environmental microplastics come from a wide range of sources, including: processing and use of plastic products, wear and weathering of artificial products, microplastic particles in cosmetics, personal care products and detergents, decomposition and degradation of plastic waste, etc. These microplastics continue to accumulate in the environment, and additives in microplastics and harmful substances adsorbed on the surface of plastics can be released, causing potential harm to the environment and human health. Different types of microplastics may have different chemical compositions and physical properties, so type discrimination analysis can help further study their toxicity and ecological impacts, and at the same time, it can help monitor and evaluate the degree and distribution of microplastic pollution in the environment. By understanding the presence of different types of microplastics in the environment, we can more accurately understand their sources, migration and risk characteristics, so as to formulate corresponding environmental protection and pollution control strategies. Constructing an accurate discrimination analysis method that can simultaneously identify multiple microplastics in the environment is an important step in solving the problem of microplastic pollution.
[0003] Meta learning is a machine learning method that focuses on enabling models to quickly adapt and improve their performance when faced with new tasks. The core idea is to learn how to learn. By training on multiple related tasks, the model can learn the common features or strategies between tasks, thereby improving its generalization ability and adaptability, and achieving higher performance and effects in various application scenarios. Summary of the invention
[0004] The purpose of this application is to provide a meta-learning-based environmental microplastic classification method, equipment, medium and product to improve the accuracy of microplastic classification.
[0005] To achieve the above objectives, this application provides the following solutions:
[0006] In the first aspect, the present application provides a method for classifying environmental microplastics based on meta-learning, comprising:
[0007] Obtain infrared spectrum information of environmental microplastic samples to be classified;
[0008] Preprocessing the infrared spectrum information to obtain processed infrared spectrum information; the preprocessing includes abnormal spectrum removal, noise reduction, baseline correction and normalization processing;
[0009] Based on the processed infrared spectral information, the category and confidence of the environmental microplastic sample to be classified are determined using an environmental microplastic classification model; wherein the environmental microplastic classification model is obtained by training a deep neural network model using a meta-learning method based on a training data set; the training data set includes infrared spectral information and corresponding categories of environmental microplastic samples for training.
[0010] Optionally, a deep neural network model is trained using a meta-learning method based on a training data set, specifically including:
[0011] Get the training dataset;
[0012] Preprocessing the infrared spectrum information of the environmental microplastic sample for training to obtain processed infrared spectrum information of the environmental microplastic sample for training;
[0013] The processed infrared spectrum information of the training environmental microplastic samples is used as input, and the categories corresponding to the training environmental microplastic samples are used as output. The deep neural network model is trained using a meta-learning method to obtain an environmental microplastic classification model.
[0014] Optionally, preprocessing the infrared spectrum information to obtain processed infrared spectrum information specifically includes:
[0015] The 3σ principle is used to eliminate abnormal spectra in infrared spectrum information, and the infrared spectrum information after eliminating abnormal spectra is obtained;
[0016] The Savitzky-Golay five-point smoothing method is used to reduce the noise of the infrared spectrum information after removing the abnormal spectrum, and the infrared spectrum information after noise reduction is obtained;
[0017] The infrared spectrum information after noise reduction is baseline corrected by wavelet transform method to obtain corrected infrared spectrum information;
[0018] The corrected infrared spectrum information is normalized to obtain processed infrared spectrum information.
[0019] Optionally, the deep neural network model includes an input layer, multiple hidden layers and a softmax layer; the input layer is connected to the multiple hidden layers respectively; and the multiple hidden layers are all connected to the softmax layer.
[0020] Optionally, infrared spectrum information of the environmental microplastic sample to be classified is obtained, specifically including:
[0021] The infrared spectrum information of the environmental microplastic sample to be classified is measured using a Fourier infrared spectrometer.
[0022] In a second aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any one of the above-described meta-learning-based environmental microplastic classification methods.
[0023] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned meta-learning-based environmental microplastic classification methods.
[0024] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned meta-learning-based environmental microplastic classification methods.
[0025] According to the specific embodiments provided in this application, this application discloses the following technical effects:
[0026] The present application provides a method, device, medium and product for classifying environmental microplastics based on meta-learning, which obtains infrared spectral information of environmental microplastic samples to be classified; pre-processes the infrared spectral information to obtain processed infrared spectral information; based on the processed infrared spectral information, uses an environmental microplastic classification model to determine the category and confidence of the environmental microplastic samples to be classified; wherein the environmental microplastic classification model is obtained by training a deep neural network model based on a training data set using a meta-learning method; the training data set includes infrared spectral information and corresponding categories of environmental microplastic samples for training. In the present application, the deep neural network model is trained by a meta-learning method, and the obtained environmental microplastic classification model can accurately classify environmental microplastics. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0028] Figure 1 A schematic diagram of a process for classifying environmental microplastics based on meta-learning provided in one embodiment of the present application;
[0029] Figure 2 Schematic diagram of meta-learning training principle;
[0030] Figure 3 Schematic diagram of the confusion matrix of microplastic classification results;
[0031] Figure 4A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION
[0032] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0033] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.
[0034] In an exemplary embodiment, Figure 1 As shown, a meta-learning-based environmental microplastic classification method is provided, including the following steps:
[0035] S1: Obtaining infrared spectrum information of the environmental microplastic sample to be classified. In practical applications, a Fourier transform infrared spectrometer is used to measure the infrared spectrum information of the environmental microplastic sample to be classified.
[0036] S2: Preprocessing the infrared spectrum information to obtain processed infrared spectrum information; the preprocessing includes abnormal spectrum removal, noise reduction, baseline correction and normalization processing.
[0037] As an optional implementation, S2 specifically includes:
[0038] The 3σ principle is used to eliminate abnormal spectra in the infrared spectrum information, and the infrared spectrum information after eliminating the abnormal spectra is obtained.
[0039] The Savitzky-Golay five-point smoothing method is used to denoise the infrared spectrum information after removing the abnormal spectrum to obtain the denoised infrared spectrum information.
[0040] The wavelet transform method is used to perform baseline correction on the infrared spectrum information after noise reduction to obtain the corrected infrared spectrum information.
[0041] The corrected infrared spectrum information is normalized to obtain processed infrared spectrum information.
[0042] S3: Based on the processed infrared spectrum information, the environmental microplastic classification model is used to determine the category and confidence of the environmental microplastic sample to be classified; wherein the environmental microplastic classification model is obtained by training the deep neural network model based on the training data set using the meta-learning method; the training data set includes the infrared spectrum information and corresponding categories of the environmental microplastic samples for training. The deep neural network model includes an input layer, multiple hidden layers and a softmax layer; the input layer is respectively connected to the multiple hidden layers; and the multiple hidden layers are all connected to the softmax layer.
[0043] As an optional implementation method, the deep neural network model is trained using a meta-learning method based on the training data set, specifically including:
[0044] Get the training dataset.
[0045] The infrared spectrum information of the training environmental microplastic sample is preprocessed to obtain the processed infrared spectrum information of the training environmental microplastic sample.
[0046] The processed infrared spectrum information of the training environmental microplastic samples is used as input, and the categories corresponding to the training environmental microplastic samples are used as output. The deep neural network model is trained using a meta-learning method to obtain an environmental microplastic classification model.
[0047] In practical applications, the training process is as follows:
[0048] (1) Data preparation: Obtain environmental microplastic samples for training, use Fourier transform infrared spectrometer to measure the infrared spectrum information of the samples (environmental microplastic samples for training), and represent each sample as a feature vector. Match the obtained sample spectrum with the spectrum database, and save each group of data in a classified manner.
[0049] like Figure 2 As shown in the figure, in meta-learning, two types of tasks are usually defined: meta-task and base task. Meta-task refers to the task of learning how to learn, while base task is a specific learning task. The meta-learning model optimizes its own learning strategy by learning from the experience of a series of base tasks to adapt to unknown new tasks. Therefore, it is necessary to construct a dataset containing multiple base tasks, each of which has its own training set and test set, called support set and query set respectively. Each base task randomly selects two types of sample spectra for training.
[0050] The obtained spectral data is preprocessed, specifically:
[0051] Abnormal spectrum removal: Identify and remove abnormal spectra based on the 3σ principle, calculate the mean and standard deviation of each wavelength point, and determine the range of mean ± 3 times standard deviation (3σ). Compare each value in the spectral data with this 3σ range, and any value lower than (mean - 3σ) or higher than (mean + 3σ) is considered an abnormal value and removed.
[0052] Noise reduction: Savitzky-Golay five-point smoothing is used for noise reduction.
[0053] Baseline correction: The spectral baseline was subtracted using wavelet transform method.
[0054] Normalization: The SNV method was used to optimize the quality of spectral data, reduce the impact of irrelevant changes and noise, and improve the comparability of data and the performance of analytical models.
[0055] (2) Construct a deep neural network model: Design a deep neural network model as a base learner for meta-learning. The model includes an input layer, multiple hidden layers, and a softmax layer as the output layer for multi-classification tasks, with mean square error as the loss function.
[0056] (3) Meta-training: Taking a base task as an example, the training process is as follows: Randomly select two types of spectral data from the preprocessed total data set as a base task, and divide the support set and query set in a ratio of 8:2. Use stochastic gradient descent as the optimizer, and train the base-learner on each base task with the support set as input to obtain the classification model DNN1 and parameters of the base task. Use the query set to test the classification effect of the base-learner and calculate the loss value. Complete the training of a base task.
[0057] The training of a base task is called an episode. Three base tasks are selected at a time to form a batch. A meta-train can train multiple batches. The training is completed after traversing all batches. The trained model and parameters are fed back to the meta-learner. The meta-learner summarizes the commonalities of each base-learner, performs fast and accurate reasoning on the new task, and transmits the reasoning to the base-learner on the test task as the initial model and initial parameter value.
[0058] (4) Meta-test: Use the tasks in the meta-test set to measure the performance of the model, and select the support set in the test task to train and fine-tune the model: Use the model trained in the meta-training phase and select the stochastic gradient descent algorithm to update the model parameters, including learning rate, batch size, regularization coefficient, number of fine-tuning steps, etc. Then, use the query set to make predictions, calculate the error between the model prediction output and the true label of the query set, and evaluate the accuracy, generalization ability, and adaptability of the model on the new task.
[0059] (5) Optimization and adjustment: After the meta-test, check the performance of the model on each task, including accuracy, precision, recall, F1 score and other indicators, evaluate the generalization ability of the model on new tasks, and determine whether there is an overfitting or underfitting problem. If the evaluation result is not ideal, the model needs to be optimized and adjusted to improve the model architecture, adjust hyperparameters or other related settings to improve performance. If the requirements are met, save the model structure, parameters, data prediction results and performance indicators.
[0060] After the above process is completed, the pre-trained model (environmental microplastic classification model) obtained through meta-training and meta-testing is loaded; the spectral information of the unknown sample is obtained for pre-processing to make it consistent with the data format during model training; the pre-trained model is used to predict the spectrum of the new sample, and the classification result and the confidence of the prediction are output.
[0061] Through the above steps, the meta-learning model can be used to quickly and accurately classify environmental microplastics.
[0062] This method was used to classify 15 types of microplastics. The confusion matrix of the classification results is shown in the figure below: Figure 3 As shown, Figure 3 The horizontal axis is the predicted label, and the vertical axis is the true label. The classification accuracy of the training set reached 94.68%, the precision was 93.61%, the recall rate was 89.01%, and the F1 score was 91.52%. The model prediction effect was good.
[0063] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 4As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store spectral data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a meta-learning-based environmental microplastic classification method is implemented.
[0064] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the present application scheme, and does not constitute a limitation on the computer device to which the present application scheme is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the above-mentioned environmental microplastic classification method based on meta-learning is implemented.
[0065] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, which, when executed by a processor, implements the above-mentioned meta-learning-based environmental microplastic classification method.
[0066] In an exemplary embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the above-mentioned meta-learning-based environmental microplastic classification method.
[0067] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0068] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0069] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.
[0070] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0071] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.
Claims
1. A meta-learning-based classification method for environmental microplastics, It is characterized in that include: Obtain infrared spectrum information of environmental microplastic samples to be classified; Preprocessing the infrared spectrum information to obtain processed infrared spectrum information; The preprocessing includes abnormal spectrum removal, noise reduction, baseline correction and normalization processing; Based on the processed infrared spectral information, the category and confidence of the environmental microplastic sample to be classified are determined using an environmental microplastic classification model; wherein the environmental microplastic classification model is obtained by training a deep neural network model using a meta-learning method based on a training data set; the training data set includes infrared spectral information and corresponding categories of environmental microplastic samples for training.
2. The method for classifying environmental microplastics based on meta-learning according to claim 1, It is characterized in that The deep neural network model is trained using meta-learning methods based on the training data set, including: Get the training dataset; Preprocessing the infrared spectrum information of the environmental microplastic sample for training to obtain processed infrared spectrum information of the environmental microplastic sample for training; The processed infrared spectrum information of the training environmental microplastic samples is used as input, and the categories corresponding to the training environmental microplastic samples are used as output. The deep neural network model is trained using a meta-learning method to obtain an environmental microplastic classification model.
3. The method for classifying environmental microplastics based on meta-learning according to claim 1 or 2, It is characterized in that Preprocessing the infrared spectrum information to obtain processed infrared spectrum information specifically includes: The 3σ principle is used to eliminate abnormal spectra in infrared spectrum information, and the infrared spectrum information after eliminating abnormal spectra is obtained; The Savitzky-Golay five-point smoothing method is used to reduce the noise of the infrared spectrum information after removing the abnormal spectrum, and the infrared spectrum information after noise reduction is obtained; The infrared spectrum information after noise reduction is baseline corrected by wavelet transform method to obtain corrected infrared spectrum information; The corrected infrared spectrum information is normalized to obtain processed infrared spectrum information.
4. The method for classifying environmental microplastics based on meta-learning according to claim 1, It is characterized in that The deep neural network model includes an input layer, multiple hidden layers and a softmax layer; the input layer is connected to the multiple hidden layers respectively; and the multiple hidden layers are all connected to the softmax layer.
5. The method for classifying environmental microplastics based on meta-learning according to claim 1, It is characterized in that Obtain infrared spectrum information of environmental microplastic samples to be classified, including: The infrared spectrum information of the environmental microplastic sample to be classified is measured using a Fourier infrared spectrometer.
6. A computer device, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the meta-learning-based environmental microplastic classification method described in any one of claims 1 to 5.
7. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, it implements the environmental microplastic classification method based on meta-learning described in any one of claims 1 to 5.
8. A computer program product comprising a computer program, It is characterized in that When the computer program is executed by a processor, it implements the environmental microplastic classification method based on meta-learning described in any one of claims 1 to 5.
Citation Information
Patent Citations
Mixed waste plastic classification method based on multi-algorithm collaborative optimization selection
CN114863286A
Hyperspectral plastic garbage material rapid classification method and device and storage medium
CN115984632A
Quick and intelligent life source waste plastic classification method based on machine learning
CN116106253A
Vehicle identification method based on distributed optical fiber strain sensing
CN116955990A
Prediction method and device for protein content in flour, spectrograph and medium
CN118329822A
