Environmental pollution case tracing method and system based on artificial intelligence

Abstract

The invention discloses an environmental pollution case tracing method and system based on artificial intelligence, and the method comprises the steps: collecting the sample data of a pollution region, the sample number comprising environmental data and a high-resolution image; preprocessing the sample data to obtain processed data; performing feature extraction on the processed data to obtain pollution features; constructing a traceability model based on the pollution characteristics; and the traceability model is used to complete environment pollution case traceability. According to the invention, by using big data and artificial intelligence technologies, rapid and accurate positioning of an environmental pollution source is realized, the efficiency and effect of environmental law enforcement are improved, and the system has wide application prospects and popularization values. The environment law enforcement officer can be effectively guided and helped to take measures rapidly, and pollution diffusion is restrained.

Description

Technical Field

[0001] The present invention relates to the field of environmental monitoring, and in particular to an artificial intelligence-based environmental pollution case tracing method and system. Background Art

[0002] Environmental monitoring refers to the continuous and systematic observation, measurement and analysis of environmental factors and their changes, aiming to timely discover and predict environmental problems and provide a scientific basis for environmental management and decision-making. With the rapid development of industrialization and urbanization, environmental pollution problems are becoming increasingly serious. Environmental problems such as air pollution, water pollution, and noise pollution pose a serious threat to human health and the ecosystem. Therefore, the research and application of environmental monitoring technology and pollution source tracing technology have become particularly important.

[0003] Traditional environmental monitoring methods rely on fixed monitoring sites, which are usually equipped with expensive and complex equipment to collect pollutant data in the air, water and soil. However, this method has disadvantages such as limited data coverage and low monitoring frequency, and it is difficult to fully and real-time reflect the dynamic process of environmental changes. With the increasing severity of environmental problems, the research and development of efficient environmental monitoring technology has important practical significance and application value. It not only helps to improve the accuracy and efficiency of environmental monitoring, but also provides strong technical support for environmental protection and governance. Summary of the invention

[0004] In order to solve the technical problems in the above background, the present invention aims to provide an environmental pollution case tracing method and system based on artificial intelligence, which can realize the rapid and accurate positioning of environmental pollution sources by utilizing big data and artificial intelligence technology, thereby improving the efficiency of environmental law enforcement and protecting the ecological environment.

[0005] To achieve the above purpose, the present invention provides an artificial intelligence-based environmental pollution case tracing method, the steps comprising:

[0006] Collecting sample data of the polluted area, the sample data includes: environmental data and high-resolution images;

[0007] Preprocessing the sample data to obtain processed data;

[0008] Performing feature extraction on the processed data to obtain pollution features;

[0009] Based on the pollution characteristics, a source tracing model is constructed;

[0010] The traceability model is used to complete the traceability of environmental pollution cases.

[0011] Preferably, the method for performing the preprocessing includes: cleaning and standardizing the collected sample data to remove noise and outliers; and then integrating multi-source data through data fusion technology to improve the integrity and accuracy of the data.

[0012] Preferably, the method for extracting features comprises:

[0013] Based on the processed data, valuable features are extracted from different data sources;

[0014] The extracted features are subjected to feature fusion to obtain the pollution features.

[0015] Preferably, for high-resolution image data, a local binary pattern is used to extract distinguishing features from the image data. Specifically, for each pixel (x c ,y c ), define a circular neighborhood with a radius of R, and there are P pixels in this neighborhood, whose coordinates are (x i ,y i ), where i=0, 1, ..., P-1;

[0016] Compare the gray value of each pixel with the central pixel I(x c ,y c ) for comparison, marking pixels with higher grayscale values ​​than the center pixel as 1 and pixels with lower grayscale values ​​as 0. After the comparison is completed, a binary string consisting of P binary values ​​is obtained, which is called the local binary pattern LBP. P，R (x c ,y c ):

[0017]

[0018] Among them, LBP P，R (x c ,y c ) means (x c ,y c ) is the local binary pattern centered at I(x i ,y i ) represents the gray value of the i-th pixel in the neighborhood; I(x c ,y c ) represents the center pixel; s(·) represents the sign function, which is defined as:

[0019]

[0020] Preferably, SVM is used to perform the feature fusion:

[0021] f(x)=sign(<w,x> +b)

[0022] In the formula, sign represents the sign function, which is used to convert the predicted value into a binary category; x represents the input feature vector; w represents the weight vector, which determines the direction of the hyperplane; b represents the bias term, which is used to adjust the position of the decision boundary;<w,x> Represents the inner product operation, which represents the dot product of the weight vector w and the input vector x.

[0023] Preferably, the optimization step of SVM includes:

[0024]

[0025] In the formula, min w，b Indicates minimization of variables w and b; ||w|| 2 represents the square of the Euclidean norm of the weight vector; C represents the regularization parameter; y i represents the label of the i-th training sample; max(0, 1-y i ( <w,x i >+b)) represents a soft margin, which allows some data points to be misclassified but will be penalized by the regularization term.

[0026] The present invention also provides an artificial intelligence-based environmental pollution case tracing system, which is used to implement the above method and includes: a collection module, a preprocessing module, an extraction module, a construction module and a tracing module;

[0027] The acquisition module is used to collect sample data of the polluted area, and the sample data includes: environmental data and high-resolution images;

[0028] The preprocessing module is used to preprocess the sample data to obtain processed data;

[0029] The extraction module is used to extract features from the processed data to obtain pollution features;

[0030] The construction module is used to construct a source tracing model based on the pollution characteristics;

[0031] The traceability module is used to complete the traceability of environmental pollution cases using the traceability model.

[0032] Preferably, the workflow of the preprocessing module includes: cleaning and standardizing the collected sample data to remove noise and outliers; then, integrating multi-source data through data fusion technology to improve data integrity and accuracy.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] The present invention uses big data and artificial intelligence technology to quickly and accurately locate environmental pollution sources, improve the efficiency and effectiveness of environmental law enforcement, and has broad application prospects and promotion value. The present invention can effectively guide environmental law enforcement personnel and help them take quick measures to curb the spread of pollution. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0036] Figure 1 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0039] Embodiment 1

[0040] like Figure 1 FIG. 1 is a schematic diagram of the method flow of this embodiment, and the steps include:

[0041] S1. Collect sample data from polluted areas.

[0042] In this embodiment, the sample data includes: environmental data and high-resolution images. The collection methods include but are not limited to sensor networks, satellite images, drone monitoring, etc. The sensor network is responsible for collecting environmental data in real time, such as air quality and water quality monitoring data; satellite images and drone monitoring provide high-resolution spatial information for locating pollution sources and monitoring pollution spread.

[0043] S2. Preprocess the sample data to obtain processed data.

[0044] After the sample data is collected, the environmental data is cleaned and preprocessed, including removing noise data, filling missing values, and unifying data formats to ensure data quality and consistency. After that, the data is aligned and integrated. Specifically, the cleaned data is aligned according to the time and space dimensions for subsequent analysis and model building; for example, ensuring that the timestamps of all data are consistent, or mapping spatial data to the same coordinate system.

[0045] For high-resolution images, wavelet denoising is used to reduce the noise in the image and improve the accuracy of subsequent feature extraction and recognition. Contrast enhancement and color correction are also used to improve the accuracy of the image. Contrast enhancement, by adjusting the gray level range of image pixels, enhances the contrast of the image, making the details in the image clearer and more prominent; color correction, adjusts the color in the image to eliminate color deviation caused by changes in light conditions, ensuring that the image is consistent under different lighting conditions. After preprocessing, processed data is obtained.

[0046] S3. Perform feature extraction on the processed data to obtain pollution features.

[0047] Based on the processed data, valuable features are extracted from different data sources and fused to enhance the information content of the data. By integrating multi-source data, the limitations of each data source are filled, thereby improving the comprehensiveness and completeness of the data. At the same time, through multi-angle data observation and analysis, the possible errors of a single data source are reduced, and the accuracy and credibility of data analysis are improved. The fused data can provide more input variables and features for the environmental pollution source tracing model, making the model more complex and comprehensive, and better able to adapt to complex environmental conditions and changes.

[0048] Specifically, the pollutant concentration and the geographical features around the pollution source are extracted through environmental data and high-resolution image data to more accurately locate the pollution source. For environmental data, data cleaning is first performed to remove noise and outliers. sco r e Perform normalization processing, specifically:

[0049]

[0050] Where X represents the original data point; X min The minimum value in the data set; X max Indicates the maximum value in the data set; X norm Represents the normalized data point. During the normalization process, if there are outliers in the data, they need to be processed before normalization to avoid excessive impact on the normalization result; at the same time, the information of the original data should be retained as much as possible during the normalization process to avoid excessive compression or loss of important features.

[0051] Then, the environmental data of the polluted area is extracted, including pollutant concentration characteristics and environmental condition characteristics. The pollutant concentration characteristics include: PM 2.5 、PM 10 , NO2, SO2, CO, O3 and other pollutant concentrations; while environmental conditions include temperature, humidity, wind speed, wind direction, etc., which affect the diffusion and deposition of pollutants.

[0052] For high-resolution image data, local binary patterns are used to extract distinguishing features from the image data. For each pixel (x c ,y c ), define a circular neighborhood with a radius of R, and there are P pixels in this neighborhood, whose coordinates are (x i ,y i ), where i=0, 1, ..., P-1.

[0053] Compare the gray value of each pixel with the central pixel I(x c ,y c ) for comparison, marking pixels with higher grayscale values ​​than the center pixel as 1 and pixels with lower grayscale values ​​as 0. After the comparison is completed, a binary string consisting of P binary values ​​is obtained, which is called the local binary pattern LBP. P，R (x c ,y c ):

[0054]

[0055] Among them, LBP P，R (x c ,y c ) means (x c ,y c ) is the local binary pattern centered at I(x i ,y i ) represents the gray value of the i-th pixel in the neighborhood; I(x c ,y c ) represents the center pixel; s(·) represents the sign function, which is defined as:

[0056]

[0057] After feature extraction is completed, a model-driven feature fusion method is used to fuse multi-source data features to obtain pollution features. The model-driven data fusion method aims to learn and fuse information from multiple data sources through a machine learning model to enhance the information content and representation ability of the data; this embodiment uses a support vector machine to perform feature fusion.

[0058] Support vector machine (SVM) is a powerful supervised learning model that is widely used in classification and regression analysis problems. In environmental pollution case tracing, support vector machine can be used to integrate information from different data sources, such as air quality data, satellite image features, etc., to predict the location of the pollution source.

[0059] f(x)=sign(<w,x> +b)

[0060] In the formula, sign represents the sign function, which is used to convert the predicted value into a binary category; x represents the input feature vector; w represents the weight vector, which determines the direction of the hyperplane; b represents the bias term, which is used to adjust the position of the decision boundary;<w,x> Represents the inner product operation, which represents the dot product of the weight vector w and the input vector x.

[0061] The goal of SVM is to find a hyperplane so that data points of different categories are separated as much as possible by the hyperplane, and the distance between the hyperplane and the nearest data point (i.e., support vector) is maximized. This method of maximizing distance can improve the generalization ability of the model. In this embodiment, the optimization problem of SVM is expressed as follows:

[0062]

[0063] In the formula, min w，b Indicates minimization of variables w and b; ||w|| 2 represents the square of the Euclidean norm of the weight vector; C represents the regularization parameter; y i represents the label of the i-th training sample; max(0, 1-y i ( <w,x i >+b)) represents a soft margin, which allows some data points to be misclassified but will be penalized by the regularization term.

[0064] S4. Based on the pollution characteristics, a traceability model is constructed.

[0065] The extracted pollution features are used to construct a source tracing model, and historical pollution data is input to train it. Specifically, a convolutional neural network (CNN) is used to construct a recognition model. In this embodiment, the recognition model includes: an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer.

[0066] The convolution layer contains three convolutions, which include 32, 64, and 128 convolution kernels respectively, and the convolution kernel size is 3×3. Multiple convolution layers learn to extract the extracted pollution features. Among them, the pooling layer size is 2×2, and the step size is 2. The number of hidden layer nodes of the fully connected layer can be adjusted according to the specific problem and the size of the data set, usually between hundreds and thousands of nodes. The softmax activation function of the output layer is used for multi-category classification to output the pollution source of the polluted area.

[0067] S5. Use the traceability model to complete the traceability of environmental pollution cases.

[0068] After the model is built, when a new polluted area appears, the environmental data and image data of the area will be input into the source tracing model to find out where the pollution source in the area is, thus achieving the effect of pollution tracing.

[0069] Embodiment 2

[0070] This embodiment also provides an artificial intelligence-based environmental pollution case tracing system, including: an acquisition module, a preprocessing module, an extraction module, a construction module and a tracing module; the acquisition module is used to collect sample data of the polluted area, and the sample data includes: environmental data and high-resolution images; the preprocessing module is used to preprocess the sample data to obtain processed data; the extraction module is used to extract features from the processed data to obtain pollution features; the construction module is used to construct a tracing model based on the pollution features; the tracing module is used to use the tracing model to complete the tracing of environmental pollution cases.

[0071] The following will describe in detail how the present invention solves the technical problems in practical work in conjunction with this embodiment.

[0072] In this embodiment, the sample data includes: environmental data and high-resolution images. The acquisition module includes but is not limited to sensor networks, satellite images, drone monitoring, etc. The sensor network is responsible for collecting environmental data in real time, such as air quality and water quality monitoring data; satellite images and drone monitoring provide high-resolution spatial information for locating the location of pollution sources and monitoring the spread of pollution.

[0073] After the sample data is collected, the preprocessing module cleans and preprocesses the environmental data, including removing noise data, filling missing values, and unifying data formats to ensure data quality and consistency. After that, data alignment and integration are performed. Specifically, the cleaned data is aligned according to time and space dimensions for subsequent analysis and model building; for example, ensuring that the timestamps of all data are consistent, or mapping spatial data to the same coordinate system.

[0074] For high-resolution images, wavelet denoising is used to reduce the noise in the image and improve the accuracy of subsequent feature extraction and recognition. Contrast enhancement and color correction are also used to improve the accuracy of the image. Contrast enhancement, by adjusting the gray level range of image pixels, enhances the contrast of the image, making the details in the image clearer and more prominent; color correction, adjusts the color in the image to eliminate color deviation caused by changes in light conditions, ensuring that the image is consistent under different lighting conditions. After preprocessing, processed data is obtained.

[0075] Based on the processed data, the extraction module extracts valuable features from different data sources and fuses them to enhance the information content of the data. By integrating multi-source data, the limitations of each data source are filled, thereby improving the comprehensiveness and completeness of the data. At the same time, through multi-angle data observation and analysis, the possible errors of a single data source are reduced, and the accuracy and credibility of data analysis are improved. The fused data can provide more input variables and features for the environmental pollution source tracing model, making the model more complex and comprehensive, and better able to adapt to complex environmental conditions and changes.

[0076] Specifically, the pollutant concentration and the geographical features around the pollution source are extracted through environmental data and high-resolution image data to more accurately locate the pollution source. For environmental data, data cleaning is first performed to remove noise and outliers. Z-score is used for normalization, specifically:

[0077]

[0078] Where X represents the original data point; X min The minimum value in the data set; X max Indicates the maximum value in the data set; X norm Represents the normalized data point. During the normalization process, if there are outliers in the data, they need to be processed before normalization to avoid excessive impact on the normalization result; at the same time, the information of the original data should be retained as much as possible during the normalization process to avoid excessive compression or loss of important features.

[0079] Then, the environmental data of the polluted area is extracted, including pollutant concentration characteristics and environmental condition characteristics. The pollutant concentration characteristics include: PM 2.5 、PM 10 , NO2, SO2, CO, O3 and other pollutant concentrations; while environmental conditions include temperature, humidity, wind speed, wind direction, etc., which affect the diffusion and deposition of pollutants.

[0080] For high-resolution image data, local binary patterns are used to extract distinguishing features from the image data. For each pixel (x c ,y c ), define a circular neighborhood with a radius of R, and there are P pixels in this neighborhood, whose coordinates are (x i ,y i ), where i=0, 1, ..., P-1.

[0081] Compare the gray value of each pixel with the central pixel I(x c ,y c ) for comparison, marking pixels with higher grayscale values ​​than the center pixel as 1 and pixels with lower grayscale values ​​as 0. After the comparison is completed, a binary string consisting of P binary values ​​is obtained, which is called the local binary pattern LBP. P，R (x c ,y c ):

[0082]

[0083] Among them, LBP P，R (x c ,y c ) means (x c ,y c ) is the local binary pattern centered at I(x i ,y i ) represents the gray value of the i-th pixel in the neighborhood; I(x c ,y c ) represents the center pixel; s(·) represents the sign function, which is defined as:

[0084]

[0085] After feature extraction is completed, a model-driven feature fusion method is used to fuse multi-source data features to obtain pollution features. The model-driven data fusion method aims to learn and fuse information from multiple data sources through a machine learning model to enhance the information content and representation ability of the data; this embodiment uses a support vector machine to perform feature fusion.

[0086] Support vector machine (SVM) is a powerful supervised learning model that is widely used in classification and regression analysis problems. In environmental pollution case tracing, support vector machine can be used to integrate information from different data sources, such as air quality data, satellite image features, etc., to predict the location of the pollution source.

[0087] f(x)=sign(<w,x> +b)

[0088] In the formula, sign represents the sign function, which is used to convert the predicted value into a binary category; x represents the input feature vector; w represents the weight vector, which determines the direction of the hyperplane; b represents the bias term, which is used to adjust the position of the decision boundary;<w,x> Represents the inner product operation, which represents the dot product of the weight vector w and the input vector x.

[0089] The goal of SVM is to find a hyperplane so that data points of different categories are separated as much as possible by the hyperplane, and the distance between the hyperplane and the nearest data point (i.e., support vector) is maximized. This method of maximizing distance can improve the generalization ability of the model. In this embodiment, the optimization problem of SVM is expressed as follows:

[0090]

[0091] In the formula, min w，b Indicates minimization of variables w and b; ||w|| 2 represents the square of the Euclidean norm of the weight vector; C represents the regularization parameter; y i represents the label of the i-th training sample; max(0, 1-y i ( <w,x i >+b)) represents a soft margin, which allows some data points to be misclassified but will be penalized by the regularization term.

[0092] The construction module uses the extracted pollution features to construct a traceability model and inputs historical pollution data to train it. Specifically, a convolutional neural network (CNN) is used to construct a recognition model. In this embodiment, the recognition model includes: an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer.

[0093] The convolution layer contains three convolutions, which include 32, 64, and 128 convolution kernels respectively, and the convolution kernel size is 3×3. Multiple convolution layers learn to extract the extracted pollution features. Among them, the pooling layer size is 2×2, and the step size is 2. The number of hidden layer nodes of the fully connected layer can be adjusted according to the specific problem and the size of the data set, usually between hundreds and thousands of nodes. The softmax activation function of the output layer is used for multi-category classification to output the pollution source of the polluted area.

[0094] Finally, the source tracing model is used to complete the source tracing of environmental pollution cases. After the model is built, when a new polluted area appears, the environmental data and image data of the area are input into the source tracing model, and the source of pollution in the area can be obtained, achieving the effect of pollution source tracing.

[0095] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for tracing the source of environmental pollution cases based on artificial intelligence, characterized in that the steps include: Collecting sample data of the polluted area, the sample data includes: environmental data and high-resolution images; Preprocessing the sample data to obtain processed data; Performing feature extraction on the processed data to obtain pollution features; Based on the pollution characteristics, a source tracing model is constructed; The traceability model is used to complete the traceability of environmental pollution cases.

2. The method for tracing the source of environmental pollution cases based on artificial intelligence according to claim 1 is characterized in that: The method for performing the preprocessing includes: cleaning and standardizing the collected sample data to remove noise and outliers; and then integrating multi-source data through data fusion technology to improve the integrity and accuracy of the data.

3. The method for tracing the source of environmental pollution cases based on artificial intelligence according to claim 1 is characterized in that: The methods for feature extraction include: Based on the processed data, valuable features are extracted from different data sources; The extracted features are subjected to feature fusion to obtain the pollution features.

4. The method for tracing the source of environmental pollution cases based on artificial intelligence according to claim 3 is characterized in that: For high-resolution image data, local binary patterns are used to extract distinguishing features from the image data. Specifically, for each pixel (x c ,y c ), define a circular neighborhood with a radius of R, and there are P pixels in this neighborhood, whose coordinates are (x i ,y i ), where i = 0, 1, ..., P-1; Compare the gray value of each pixel with the central pixel I(x c ,y c ) is used for comparison, and pixels with higher grayscale values ​​than the center pixel are marked as 1, and pixels with lower grayscale values ​​are marked as 0. After the comparison is completed, a binary string consisting of P binary values ​​is obtained, which is called the local binary pattern LBP. P，R (x c ,y c ): Among them, LBP P，R (x c ,y c ) means (x c ,y c ) is the local binary pattern centered at I(x i ,y i ) represents the gray value of the i-th pixel in the neighborhood; I(x c ,y c ) represents the center pixel; s(·) represents the sign function, which is defined as:

5. The method for tracing the source of environmental pollution cases based on artificial intelligence according to claim 3 is characterized in that: SVM is used to perform the feature fusion: f(x)=sign(<w,r> +b) In the formula, sign represents the sign function, which is used to convert the predicted value into a binary category; x represents the input feature vector; w represents the weight vector, which determines the direction of the hyperplane; b represents the bias term, which is used to adjust the position of the decision boundary;<w,x> Represents the inner product operation, which represents the dot product of the weight vector w and the input vector x.

6. The method for tracing the source of environmental pollution cases based on artificial intelligence according to claim 5 is characterized in that: The optimization steps of SVM include: In the formula, min w，b Indicates minimization of variables w and b; ||w|| 2 represents the square of the Euclidean norm of the weight vector; C represents the regularization parameter; yi represents the label of the i-th training sample; max(0, 1-y i ( <w,x i >+b)) represents a soft margin, which allows some data points to be misclassified but will be penalized by the regularization term.

7. An artificial intelligence-based environmental pollution case tracing system, the system is used to implement the method described in any one of claims 1 to 6, characterized in that: include: Acquisition module, preprocessing module, extraction module, construction module and traceability module; The acquisition module is used to collect sample data of the polluted area, and the sample data includes: environmental data and high-resolution images; The preprocessing module is used to preprocess the sample data to obtain processed data; The extraction module is used to extract features from the processed data to obtain pollution features; The construction module is used to construct a source tracing model based on the pollution characteristics; The traceability module is used to complete the traceability of environmental pollution cases using the traceability model.

8. The artificial intelligence-based environmental pollution case tracing system according to claim 7 is characterized in that: The workflow of the preprocessing module includes: cleaning and standardizing the collected sample data to remove noise and outliers; then, integrating multi-source data through data fusion technology to improve the integrity and accuracy of the data.

Images