CN116186566B - Diffusion prediction method and system based on deep learning - Google Patents
Diffusion prediction method and system based on deep learning Download PDFInfo
- Publication number
- CN116186566B CN116186566B CN202310465760.9A CN202310465760A CN116186566B CN 116186566 B CN116186566 B CN 116186566B CN 202310465760 A CN202310465760 A CN 202310465760A CN 116186566 B CN116186566 B CN 116186566B
- Authority
- CN
- China
- Prior art keywords
- target
- diffusion
- pollution
- acquiring
- target pollutant
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A20/00—Water conservation; Efficient water supply; Efficient water use
- Y02A20/152—Water filtration
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Development Economics (AREA)
- Software Systems (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Educational Administration (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Medical Informatics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Game Theory and Decision Science (AREA)
- Mathematical Physics (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Processing Of Solid Wastes (AREA)
Abstract
The invention discloses a diffusion prediction method and a diffusion prediction system based on deep learning, wherein the diffusion prediction method comprises the following steps: acquiring pollution source information in a target area, screening target pollutants, and determining target pollutant monitoring points according to GIS geographic data of the target area; acquiring target pollutant concentration information through target pollutant monitoring points to acquire a pollution diffusion pattern of a target area; acquiring a concentration change sequence of target pollutants in a pollution area, constructing a diffusion prediction model, and acquiring space-time correlation characteristics of the concentration of the target pollutants, soil properties and water quality properties in a continuous time period through the diffusion prediction model; and acquiring a potential pollution area by combining the pollution distribution of the target area at the current time with the diffusion trend of the target pollutant, and generating a predicted pollution diffusion distribution. According to the invention, the distribution condition of pollution diffusion is simulated by learning the distribution rule and the time-space correlation characteristic of the pollution, so that the accuracy of pollution diffusion prediction is improved, and a data basis is provided for pollutant prevention and control work.
Description
Technical Field
The invention relates to the technical field of pollution monitoring, in particular to a diffusion prediction method and a diffusion prediction system based on deep learning.
Background
Environmental monitoring refers to the process of monitoring, analyzing, evaluating and predicting the condition of various pollutants and their interactions in the environment. The acceleration and expansion of the urban and industrialized processes causes a great deal of pollutants to be released and spread into the atmosphere, water and soil, leading to increasingly serious pollution problems. Pollution diffusion can be classified into various types such as aerosol diffusion, water quality diffusion, soil diffusion, etc. The diffusion of the pollutants has a great influence on the surrounding environment and human health, and corresponding measures need to be taken to prevent and control the diffusion of the pollutants. Common sources of pollution include industrial waste gas, traffic exhaust, agricultural pollution, landfill, and the like.
The key to pollution diffusion prediction is to build an accurate mathematical model. Mathematical models typically include physical models, chemical models, numerical models, and the like. Physical models are used to describe the physical processes of contaminants in the atmosphere, bodies of water or soil, such as convection, diffusion, sedimentation, reflection, etc.; the chemical model is used for describing the chemical reaction and conversion process of the pollutants; the numerical model then converts the physical and chemical models into a computer program for modeling and predicting the laws of transport and diffusion of the contaminant. At present, in pollution diffusion prediction, deep learning can be used for extracting complex relation and mode of pollutant transmission and diffusion, so that accurate prediction is realized. Therefore, how to learn the transmission rule of the pollutant and the environmental factors through deep learning in pollution diffusion prediction, so as to realize accurate prediction of diffusion is an urgent problem which cannot be solved.
Disclosure of Invention
In order to solve the technical problems, the invention provides a diffusion prediction method and a diffusion prediction system based on deep learning.
The first aspect of the present invention provides a diffusion prediction method based on deep learning, comprising:
acquiring pollution source information in a target area, screening target pollutants according to the pollution source information, and determining target pollutant monitoring points according to GIS geographic data of the target area;
collecting target pollutant concentration information through target pollutant monitoring points, and clustering all monitoring points based on the target pollutant concentration information to obtain a pollution diffusion pattern of a target area;
acquiring a concentration change sequence of target pollutants in a pollution area according to a pollution diffusion pattern of the target area, constructing a diffusion prediction model, and training based on the concentration change sequence of the target pollutants in the pollution area;
acquiring space-time correlation characteristics of the concentration of the target pollutant, the soil property and the water quality property in a continuous time period through a diffusion prediction model, and representing the diffusion trend of the target pollutant through the space-time correlation characteristics;
and acquiring a potential pollution area by combining the pollution distribution of the current time of the target area with the diffusion trend of the target pollutant, generating predicted pollution diffusion distribution, and carrying out visual display.
In the scheme, pollution source information in a target area is acquired, target pollutants are screened according to the pollution source information, and target pollutant monitoring points are determined according to GIS geographic data of the target area, specifically:
acquiring pollution source information in a range according to the regional range of the target region, extracting pollution source position information and pollutants produced by a pollution source according to the pollution source information, and screening out marked pollutants of the target region through the pollutants;
screening in the marked pollutants according to a preset environment monitoring task to obtain target pollutants, obtaining position information of an existing environment monitoring well of a target area, extracting environment monitoring information in the environment monitoring well, and analyzing the existing monitoring data of the target pollutants;
carrying out data preprocessing on the existing monitoring data of the target pollutants, carrying out data analysis on the preprocessed existing monitoring data, retrieving and acquiring experience related to the diffusion of the target pollutants by a big data means, and setting a monitoring data correlation standard under the current environmental monitoring well position distribution;
acquiring data association of environmental monitoring wells at different positions according to data analysis, judging the deviation of the data association and a monitoring data correlation standard, and acquiring GIS geographic data of a target area when the deviation is not in a preset deviation interval;
And acquiring the terrain and hydrologic data in the target area according to the GIS geographic data, carrying out position optimization and newly adding monitoring points by utilizing a genetic algorithm on the basis of the existing environment monitoring well according to the pollution source position information, and generating target pollutant monitoring points by combining the environment monitoring well.
In the scheme, target pollutant concentration information is acquired through target pollutant monitoring points, each monitoring point is clustered based on the target pollutant concentration information, and a pollution diffusion pattern of a target area is acquired, specifically:
dividing a target area into grids according to the position information of target pollutant monitoring points, acquiring target pollutant concentration information acquired by the target pollutant monitoring points, and acquiring a clustering sample set by combining the grids with the target pollutant concentration information;
obtaining a similarity matrix between grid samples according to grid position information and target pollutant concentration information in a cluster sample set, obtaining attraction and attribution between the grid samples through the similarity matrix, and obtaining a cluster center of the cluster sample set after multiple iterations;
obtaining a clustering result through a clustering center of the clustering sample set, obtaining pollution distribution of a target area according to the clustering result, and determining a diffusion direction according to intra-class distances of all grid samples in the clustering result;
And generating a diffusion pattern of the target pollutant in the target area according to the pollution distribution and the diffusion direction of the target area.
In the scheme, the space-time correlation characteristics are obtained for the concentration of the target pollutant, the soil property and the water quality property in the continuous time period through the diffusion prediction model, and specifically:
acquiring a concentration change sequence of the target pollutant in the polluted area according to the historical monitoring data of the target pollutant and the diffusion pattern of the target pollutant, preprocessing the concentration change sequence, and removing abnormal monitoring data;
constructing a diffusion prediction model based on deep learning, taking the pretreated target pollutant concentration change sequence as a model input, and carrying out normalization treatment;
learning and representing a graph structure formed by target pollutant monitoring points in a target area through a graph attention network, constructing an adjacent matrix to acquire a spatial relationship between adjacent target pollutant monitoring points, and distributing different weights to neighbor nodes by combining average pooling and maximum pooling;
in addition, attention coefficients are set among target pollutant monitoring stations according to the diffusion directions, aggregation of neighbor nodes is carried out according to the attention coefficients, useless nodes are filtered, and spatial features corresponding to the target pollutant concentration change sequences are obtained according to a neighbor aggregation mechanism of a graph attention network;
Inputting a characteristic sequence output by a graph attention network into a DASEQ2Seq model for characteristic coding, acquiring the hidden state of the current timestamp by utilizing the characteristic input of the current timestamp and the hidden state of the previous timestamp, introducing time attention into the DASEQ2Seq model, and representing the importance of each characteristic in the characteristic sequence;
and after decoding by a decoder, acquiring the space-time correlation characteristic.
In the scheme, a graph structure formed by target pollutant monitoring points in a target area is learned and represented through a graph attention network, and the method specifically comprises the following steps:
acquiring position information of target pollutant monitoring points in a target area, and determining a spatial topological relationDefined as，/>Representing target contaminant monitoring points as nodes in the graph structure, +.>Representing a spatial relationship between target contaminant monitoring points;
acquiring a grid where target pollutant monitoring points are located, taking corresponding target pollutant monitoring data as node characteristics, and acquiring soil characteristics and water characteristics in the grid as additional characteristics of the nodes;
and performing representation learning on the node characteristics through a message passing mechanism and a neighbor aggregation mechanism of the graph attention network, and generating a corresponding low-dimensional vector representation.
In the scheme, the potential pollution area is obtained by combining the pollution distribution of the current time of the target area with the diffusion trend of the target pollutant, the predicted pollution diffusion distribution is generated, and the visual display is carried out, specifically:
acquiring current pollution distribution of the target pollutant according to a pollution diffusion pattern in the target area, and inputting the current pollution distribution into a trained diffusion prediction model;
extracting space-time correlation characteristics through the diffusion prediction model, and acquiring target pollutant prediction concentrations of target pollutant monitoring points after preset time according to the space-time correlation characteristics;
performing cluster analysis according to the target pollutant predicted concentration to generate a new cluster result, and comparing the new cluster result with a historical cluster result to obtain a cluster deviation;
and determining a corresponding grid according to the clustering deviation, marking, generating predicted pollution diffusion distribution by using the marked grid potential pollution area, generating related early warning information according to the predicted pollution diffusion distribution, and carrying out visual processing on the early warning information and the predicted pollution diffusion distribution.
The second aspect of the present invention also provides a diffusion prediction system based on deep learning, the system comprising: the diffusion prediction method based on the deep learning comprises a memory and a processor, wherein the memory comprises a diffusion prediction method program based on the deep learning, and the diffusion prediction method program based on the deep learning realizes the following steps when being executed by the processor:
Acquiring pollution source information in a target area, screening target pollutants according to the pollution source information, and determining target pollutant monitoring points according to GIS geographic data of the target area;
collecting target pollutant concentration information through target pollutant monitoring points, and clustering all monitoring points based on the target pollutant concentration information to obtain a pollution diffusion pattern of a target area;
acquiring a concentration change sequence of target pollutants in a pollution area according to a pollution diffusion pattern of the target area, constructing a diffusion prediction model, and training based on the concentration change sequence of the target pollutants in the pollution area;
acquiring space-time correlation characteristics of the concentration of the target pollutant, the soil property and the water quality property in a continuous time period through a diffusion prediction model, and representing the diffusion trend of the target pollutant through the space-time correlation characteristics;
and acquiring a potential pollution area by combining the pollution distribution of the current time of the target area with the diffusion trend of the target pollutant, generating predicted pollution diffusion distribution, and carrying out visual display.
The invention discloses a diffusion prediction method and a diffusion prediction system based on deep learning, wherein the diffusion prediction method comprises the following steps: acquiring pollution source information in a target area, screening target pollutants, and determining target pollutant monitoring points according to GIS geographic data of the target area; acquiring target pollutant concentration information through target pollutant monitoring points to acquire a pollution diffusion pattern of a target area; acquiring a concentration change sequence of target pollutants in a pollution area, constructing a diffusion prediction model, acquiring space-time correlation characteristics of the concentration of the target pollutants, soil properties and water quality properties in a continuous time period through the diffusion prediction model, and representing the diffusion trend of the target pollutants; and acquiring a potential pollution area by combining the pollution distribution of the target area at the current time with the diffusion trend of the target pollutant, and generating a predicted pollution diffusion distribution. According to the invention, the distribution condition of pollution diffusion is simulated by learning the distribution rule and the time-space correlation characteristic of the pollution, so that the accuracy of pollution diffusion prediction is improved, and a data basis is provided for pollutant prevention and control work.
Drawings
FIG. 1 shows a flow chart of a diffusion prediction method based on deep learning of the present invention;
FIG. 2 is a flow chart of a method of the present invention for acquiring a contamination diffusion profile of a target area;
FIG. 3 is a flow chart of a method for acquiring space-time correlation features through a diffusion prediction model according to the present invention;
FIG. 4 shows a block diagram of a diffusion prediction system based on deep learning of the present invention.
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, in the case of no conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced in other ways than those described herein, and therefore the scope of the present invention is not limited to the specific embodiments disclosed below.
FIG. 1 shows a flow chart of a diffusion prediction method based on deep learning of the present invention.
As shown in fig. 1, a first aspect of the present invention provides a diffusion prediction method based on deep learning, including:
S102, obtaining pollution source information in a target area, screening target pollutants according to the pollution source information, and determining target pollutant monitoring points according to GIS geographic data of the target area;
s104, collecting target pollutant concentration information through target pollutant monitoring points, and clustering all monitoring points based on the target pollutant concentration information to obtain a pollution diffusion pattern of a target area;
s106, acquiring a concentration change sequence of target pollutants in the pollution area according to a pollution diffusion pattern of the target area, constructing a diffusion prediction model, and training based on the concentration change sequence of the target pollutants in the pollution area;
s108, acquiring space-time correlation characteristics of the concentration of the target pollutant, the soil property and the water quality property in a continuous time period through a diffusion prediction model, and representing the diffusion trend of the target pollutant through the space-time correlation characteristics;
s110, the potential pollution area is obtained by combining the pollution distribution of the current time of the target area with the diffusion trend of the target pollutant, the predicted pollution diffusion distribution is generated, and the visual display is carried out.
The method comprises the steps of acquiring pollution source information in a range according to the regional range of a target region, extracting pollution source position information and pollutants produced by a pollution source according to the pollution source information, wherein the pollutant sources comprise natural sources, town sources, agricultural sources and the like, and screening out marked pollutants of the target region through the pollutants; screening in the marked pollutants according to a preset environment monitoring task to obtain target pollutants, obtaining position information of an existing environment monitoring well of a target area, extracting environment monitoring information in the environment monitoring well, and analyzing the existing monitoring data of the target pollutants; carrying out data preprocessing on the existing monitoring data of the target pollutants, carrying out data analysis on the preprocessed existing monitoring data to obtain GIS geographic data of the target area, retrieving and obtaining similar areas of the target area by a big data means, obtaining experience and diffusion rules related to historical diffusion of the target pollutants, and setting a monitoring data correlation standard under the current environmental monitoring well position distribution; acquiring data association of environmental monitoring wells at different positions according to data analysis, judging the deviation of the data association and the correlation standard of monitoring data, acquiring terrain and hydrologic data in a target area according to GIS geographic data when the deviation is not in a preset deviation interval, carrying out position optimization on newly added monitoring points by utilizing a genetic algorithm according to pollution source position information on the basis of the existing environmental monitoring wells, and generating target pollutant monitoring points by combining the environmental monitoring wells.
When the existing environment monitoring well does not exist in the target area, selecting target pollutant monitoring points by a method of constructing a deep learning model through a genetic algorithm according to terrain and hydrologic data in the target area, and taking the mean square distance between the monitoring points as an fitness function; and selecting candidate points of newly added detection wells in the region according to the hydrogeologic information and pollution source position information of the target region.
FIG. 2 shows a flow chart of a method of the present invention for acquiring a contamination diffusion profile of a target area.
According to the embodiment of the invention, the target pollutant concentration information is acquired through the target pollutant monitoring points, and each monitoring point is clustered based on the target pollutant concentration information to acquire the pollution diffusion pattern of the target area, which is specifically as follows:
s202, dividing a target area into grids according to position information of target pollutant monitoring points, acquiring target pollutant concentration information acquired by the target pollutant monitoring points, and acquiring a clustering sample set by combining the grids with the target pollutant concentration information;
s204, obtaining a similarity matrix between grid samples according to grid position information and target pollutant concentration information in a cluster sample set, obtaining attraction and attribution between the grid samples through the similarity matrix, and obtaining a cluster center of the cluster sample set after multiple iterations;
S206, obtaining a clustering result through a clustering center of the clustering sample set, obtaining pollution distribution of a target area according to the clustering result, and determining a diffusion direction according to intra-class distances of all grid samples in the clustering result;
s208, generating a diffusion pattern of target pollutants in the target area according to the pollution distribution and the diffusion direction of the target area.
The similarity among the grid samples is obtained according to the grid position information and the target pollutant concentration information in the clustering sample set, and a similarity matrix is generated; acquiring a negative square Euclidean distance and calculating similarity of the negative square Euclidean distance of target pollutant concentration information corresponding to the monitoring point through the position coordinate information;wherein, the method comprises the steps of, wherein,representing grid->Similarity of->Data representing the correspondence of grid i +.>Data corresponding to the representation grid j can be coordinate information or target pollutant concentration information; the attraction degree is used for evaluating whether the grid can be used as a clustering center of a clustering result, and the attribution degree is used forEvaluating the suitability of the grids as the clustering centers of the clustering results, continuously updating the attraction degree and the attribution degree through iteration, ending iterative calculation when the iteration times exceed the set times or the change of the clustering centers is small, obtaining the clustering centers, carrying out clustering analysis according to the clustering centers to obtain corresponding clustering results, obtaining the negative square Euclidean distance of the concentration of the target pollutants corresponding to each grid in the clustering results to judge the change characteristics of the monitoring data, generating data change directions, and determining the diffusion directions of the target pollutants according to the corresponding grids, wherein the diffusion directions can be single directions or multi-direction directions.
FIG. 3 shows a flow chart of a method of the present invention for obtaining spatio-temporal correlation features through a diffusion prediction model.
According to the embodiment of the invention, the space-time correlation characteristics of the target pollutant concentration, the soil property and the water quality property in the continuous time period are obtained through the diffusion prediction model, and specifically:
s302, acquiring a concentration change sequence of target pollutants in a pollution area according to historical monitoring data of the target pollutants and a diffusion pattern of the target pollutants, preprocessing the concentration change sequence, and eliminating abnormal monitoring data;
s304, a diffusion prediction model is built based on deep learning, a preprocessed target pollutant concentration change sequence is used as a model input, and normalization processing is carried out;
s306, learning and representing a graph structure formed by target pollutant monitoring points in a target area through a graph attention network, constructing an adjacent matrix to acquire a spatial relationship between adjacent target pollutant monitoring points, and distributing different weights to neighbor nodes by utilizing the combination of average pooling and maximum pooling;
s308, additionally, attention coefficients are set among the target pollutant monitoring stations according to the diffusion directions, aggregation of neighbor nodes is carried out according to the attention coefficients, useless nodes are filtered, and spatial features corresponding to the target pollutant concentration change sequence are obtained according to a neighbor aggregation mechanism of the attention network of the graph;
S310, inputting a characteristic sequence output by a graph attention network into a DASEQ2Seq model for characteristic coding, acquiring a hidden state of a current timestamp by utilizing the characteristic input of the current timestamp and the hidden state of a previous timestamp, introducing time attention into the DASEQ2Seq model, and representing the importance of each characteristic in the characteristic sequence;
s312, after decoding by a decoder, the space-time correlation characteristics are obtained.
The method includes the steps of acquiring position information of target pollutant monitoring points in a target area and determining a spatial topological relationDefined as->，/>Representing target contaminant monitoring points as nodes in the graph structure, +.>Representing a spatial relationship between target contaminant monitoring points; acquiring a grid where target pollutant monitoring points are located, taking corresponding target pollutant monitoring data as node characteristics, and acquiring soil characteristics and water characteristics in the grid as additional characteristics of the nodes; and performing representation learning on the node characteristics through a message passing mechanism and a neighbor aggregation mechanism of the graph attention network, and generating a corresponding low-dimensional vector representation.
Assigning different weights to neighbor nodes using average pooling and maximum pooling combinationsWherein the weight->The calculation formula of (2) is as follows: / >Wherein (1)>Representing an activation function->Representing a full connection layer, ">Representing a mapping function->Node characteristic representing time t,/>Representing average pooling>Representing maximum pooling;
and setting attention coefficients according to the diffusion directions among the target pollutant monitoring stations, carrying out aggregation of neighbor nodes according to the attention coefficients, when the target pollutant monitoring point i is in the diffusion direction of the target pollutant monitoring point j, setting the attention coefficient of the target pollutant monitoring point j to the target pollutant monitoring point i as 1, setting the attention coefficient of the target pollutant monitoring point i to the target pollutant monitoring point j as 0, and setting the attention coefficient to be 0 when the target pollutant monitoring point i is not in the diffusion direction, so that only useful nodes are aggregated in a neighbor aggregation mechanism, setting a multi-head attention mechanism in an attention network, respectively calculating corresponding results through a plurality of groups of mutually independent attention mechanisms, splicing the results to obtain the convolved results, and obtaining the corresponding characteristic representation.
The DASEQ2Seq model uses RNN units, generally LSTM and GRU, calculates the correlation through the hidden state and the memory state of the encoder at the last moment, normalizes through a softMax function, and obtains the corresponding time attention ，/>Wherein->Is time ofAttention weight, representing the importance of the kth feature in the sequence of features at time t to the input, +.>For the kth feature at time t, the correlation corresponding to the ith feature, ++>Is the total number of features, and is output after being decoded by a decoder through shared parameters.
The current pollution distribution of the target pollutant is obtained according to the pollution diffusion pattern in the target area, and the current pollution distribution is input into a diffusion prediction model after training; extracting space-time correlation characteristics through the diffusion prediction model, and acquiring target pollutant prediction concentrations of target pollutant monitoring points after preset time according to the space-time correlation characteristics; performing cluster analysis according to the target pollutant predicted concentration to generate a new cluster result, and comparing the new cluster result with a historical cluster result to obtain a cluster deviation; and determining a corresponding grid according to the clustering deviation, marking, generating predicted pollution diffusion distribution by using the marked grid potential pollution area, generating related early warning information according to the predicted pollution diffusion distribution, and carrying out visual processing on the early warning information and the predicted pollution diffusion distribution.
According to the embodiment of the invention, the marker plants of the target pollutants are selected for diffusion monitoring, specifically:
Acquiring target pollutant information and plant coverage information of a target area, and establishing a retrieval task according to the target pollutant information and the climate characteristics of the target area;
acquiring plant information with pollutant enrichment capacity corresponding to target pollutants in a preset search space by utilizing a big data means, optimizing the searched plant information according to plant coverage information and climate characteristics of a target area, and acquiring a plant information set meeting the requirements;
dividing the grids into similar areas according to the plant coverage information, dividing the grids with the plant coverage information similarity larger than a preset standard into the same area, and selecting the marker plants corresponding to each area from the plant information set meeting the requirements;
the method comprises the steps of obtaining the enrichment amount of pollutants in the marker plants in different periods according to the seasonal characteristics of a target area, and monitoring the diffusion of the pollutants in the same area in different periods.
FIG. 4 shows a block diagram of a diffusion prediction system based on deep learning of the present invention.
The second aspect of the present invention also provides a diffusion prediction system 4 based on deep learning, the system comprising: a memory 41, a processor 42, wherein the memory includes a diffusion prediction method program based on deep learning, and the diffusion prediction method program based on deep learning realizes the following steps when executed by the processor:
Acquiring pollution source information in a target area, screening target pollutants according to the pollution source information, and determining target pollutant monitoring points according to GIS geographic data of the target area;
collecting target pollutant concentration information through target pollutant monitoring points, and clustering all monitoring points based on the target pollutant concentration information to obtain a pollution diffusion pattern of a target area;
acquiring a concentration change sequence of target pollutants in a pollution area according to a pollution diffusion pattern of the target area, constructing a diffusion prediction model, and training based on the concentration change sequence of the target pollutants in the pollution area;
acquiring space-time correlation characteristics of the concentration of the target pollutant, the soil property and the water quality property in a continuous time period through a diffusion prediction model, and representing the diffusion trend of the target pollutant through the space-time correlation characteristics;
and acquiring a potential pollution area by combining the pollution distribution of the current time of the target area with the diffusion trend of the target pollutant, generating predicted pollution diffusion distribution, and carrying out visual display.
The method comprises the steps of acquiring pollution source information in a range according to the regional range of a target region, extracting pollution source position information and pollutants produced by a pollution source according to the pollution source information, wherein the pollutant sources comprise natural sources, town sources, agricultural sources and the like, and screening out marked pollutants of the target region through the pollutants; screening in the marked pollutants according to a preset environment monitoring task to obtain target pollutants, obtaining position information of an existing environment monitoring well of a target area, extracting environment monitoring information in the environment monitoring well, and analyzing the existing monitoring data of the target pollutants; carrying out data preprocessing on the existing monitoring data of the target pollutants, carrying out data analysis on the preprocessed existing monitoring data to obtain GIS geographic data of the target area, retrieving and obtaining similar areas of the target area by a big data means, obtaining experience and diffusion rules related to historical diffusion of the target pollutants, and setting a monitoring data correlation standard under the current environmental monitoring well position distribution; acquiring data association of environmental monitoring wells at different positions according to data analysis, judging the deviation of the data association and the correlation standard of monitoring data, acquiring terrain and hydrologic data in a target area according to GIS geographic data when the deviation is not in a preset deviation interval, carrying out position optimization on newly added monitoring points by utilizing a genetic algorithm according to pollution source position information on the basis of the existing environmental monitoring wells, and generating target pollutant monitoring points by combining the environmental monitoring wells.
When the existing environment monitoring well does not exist in the target area, selecting target pollutant monitoring points by a method of constructing a deep learning model through a genetic algorithm according to terrain and hydrologic data in the target area, and taking the mean square distance between the monitoring points as an fitness function; and selecting candidate points of newly added detection wells in the region according to the hydrogeologic information and pollution source position information of the target region.
According to the embodiment of the invention, the target pollutant concentration information is acquired through the target pollutant monitoring points, and each monitoring point is clustered based on the target pollutant concentration information to acquire the pollution diffusion pattern of the target area, which is specifically as follows:
dividing a target area into grids according to the position information of target pollutant monitoring points, acquiring target pollutant concentration information acquired by the target pollutant monitoring points, and acquiring a clustering sample set by combining the grids with the target pollutant concentration information;
obtaining a similarity matrix between grid samples according to grid position information and target pollutant concentration information in a cluster sample set, obtaining attraction and attribution between the grid samples through the similarity matrix, and obtaining a cluster center of the cluster sample set after multiple iterations;
Obtaining a clustering result through a clustering center of the clustering sample set, obtaining pollution distribution of a target area according to the clustering result, and determining a diffusion direction according to intra-class distances of all grid samples in the clustering result;
and generating a diffusion pattern of the target pollutant in the target area according to the pollution distribution and the diffusion direction of the target area.
The similarity among the grid samples is obtained according to the grid position information and the target pollutant concentration information in the clustering sample set, and a similarity matrix is generated; acquiring a negative square Euclidean distance and calculating similarity of the negative square Euclidean distance of target pollutant concentration information corresponding to the monitoring point through the position coordinate information;wherein, the method comprises the steps of, wherein,representing grid->Similarity of->Data representing the correspondence of grid i +.>Data corresponding to the representation grid j can be coordinate information or target pollutant concentration information; the attraction degree is used for evaluating whether the grid can be used as a clustering center of a clustering result, the attribution degree is used for evaluating the suitability of the grid as the clustering center of the clustering result, the attraction degree and the attribution degree are updated continuously through iteration, and when the attraction degree and the attribution degree are overlappedEnding iterative computation when the generation times exceed the set times or the change of the clustering center is small, obtaining the clustering center, carrying out clustering analysis according to the clustering center to obtain a corresponding clustering result, obtaining the negative square Euclidean distance of the concentration of the target pollutant corresponding to each grid in the clustering result to judge the change characteristics of the monitoring data, generating the change direction of the data, and determining the diffusion direction of the target pollutant according to the corresponding grids, wherein the diffusion direction can be a single direction or a multi-direction.
According to the embodiment of the invention, the space-time correlation characteristics of the target pollutant concentration, the soil property and the water quality property in the continuous time period are obtained through the diffusion prediction model, and specifically:
acquiring a concentration change sequence of the target pollutant in the polluted area according to the historical monitoring data of the target pollutant and the diffusion pattern of the target pollutant, preprocessing the concentration change sequence, and removing abnormal monitoring data;
constructing a diffusion prediction model based on deep learning, taking the pretreated target pollutant concentration change sequence as a model input, and carrying out normalization treatment;
learning and representing a graph structure formed by target pollutant monitoring points in a target area through a graph attention network, constructing an adjacent matrix to acquire a spatial relationship between adjacent target pollutant monitoring points, and distributing different weights to neighbor nodes by combining average pooling and maximum pooling;
in addition, attention coefficients are set among target pollutant monitoring stations according to the diffusion directions, aggregation of neighbor nodes is carried out according to the attention coefficients, useless nodes are filtered, and spatial features corresponding to the target pollutant concentration change sequences are obtained according to a neighbor aggregation mechanism of a graph attention network;
Inputting a characteristic sequence output by a graph attention network into a DASEQ2Seq model for characteristic coding, acquiring the hidden state of the current timestamp by utilizing the characteristic input of the current timestamp and the hidden state of the previous timestamp, introducing time attention into the DASEQ2Seq model, and representing the importance of each characteristic in the characteristic sequence;
and after decoding by a decoder, acquiring the space-time correlation characteristic.
The method includes the steps of acquiring position information of target pollutant monitoring points in a target area and determining a spatial topological relationDefined as->，/>Representing target contaminant monitoring points as nodes in the graph structure, +.>Representing a spatial relationship between target contaminant monitoring points; acquiring a grid where target pollutant monitoring points are located, taking corresponding target pollutant monitoring data as node characteristics, and acquiring soil characteristics and water characteristics in the grid as additional characteristics of the nodes; and performing representation learning on the node characteristics through a message passing mechanism and a neighbor aggregation mechanism of the graph attention network, and generating a corresponding low-dimensional vector representation.
Assigning different weights to neighbor nodes using average pooling and maximum pooling combinationsWherein the weight->The calculation formula of (2) is as follows: / >Wherein (1)>Representing an activation function->Representing a full connection layer, ">Representing a mapping function->Node characteristic representing time t,/>Representing average pooling>Representing maximum pooling;
and setting attention coefficients according to the diffusion directions among the target pollutant monitoring stations, carrying out aggregation of neighbor nodes according to the attention coefficients, when the target pollutant monitoring point i is in the diffusion direction of the target pollutant monitoring point j, setting the attention coefficient of the target pollutant monitoring point j to the target pollutant monitoring point i as 1, setting the attention coefficient of the target pollutant monitoring point i to the target pollutant monitoring point j as 0, and setting the attention coefficient to be 0 when the target pollutant monitoring point i is not in the diffusion direction, so that only useful nodes are aggregated in a neighbor aggregation mechanism, setting a multi-head attention mechanism in an attention network, respectively calculating corresponding results through a plurality of groups of mutually independent attention mechanisms, splicing the results to obtain the convolved results, and obtaining the corresponding characteristic representation.
The DASEQ2Seq model uses RNN units, generally LSTM and GRU, calculates the correlation through the hidden state and the memory state of the encoder at the last moment, normalizes through a softMax function, and obtains the corresponding time attention ，/>Wherein->For the time attention weight, the importance of the kth feature in the feature sequence at time t for the input is indicated, +.>For the kth feature at time t, the correlation corresponding to the ith feature, ++>Is the total number of features, and is output after being decoded by a decoder through shared parameters.
The current pollution distribution of the target pollutant is obtained according to the pollution diffusion pattern in the target area, and the current pollution distribution is input into a diffusion prediction model after training; extracting space-time correlation characteristics through the diffusion prediction model, and acquiring target pollutant prediction concentrations of target pollutant monitoring points after preset time according to the space-time correlation characteristics; performing cluster analysis according to the target pollutant predicted concentration to generate a new cluster result, and comparing the new cluster result with a historical cluster result to obtain a cluster deviation; and determining a corresponding grid according to the clustering deviation, marking, generating predicted pollution diffusion distribution by using the marked grid potential pollution area, generating related early warning information according to the predicted pollution diffusion distribution, and carrying out visual processing on the early warning information and the predicted pollution diffusion distribution.
The third aspect of the present invention also provides a computer-readable storage medium having embodied therein a diffusion prediction method program based on deep learning, which when executed by a processor, implements the steps of a diffusion prediction method based on deep learning as described in any one of the above.
In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above described device embodiments are only illustrative, e.g. the division of the units is only one logical function division, and there may be other divisions in practice, such as: multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. In addition, the various components shown or discussed may be coupled or directly coupled or communicatively coupled to each other via some interface, whether indirectly coupled or communicatively coupled to devices or units, whether electrically, mechanically, or otherwise.
The units described above as separate components may or may not be physically separate, and components shown as units may or may not be physical units; can be located in one place or distributed to a plurality of network units; some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present invention may be integrated in one processing unit, or each unit may be separately used as one unit, or two or more units may be integrated in one unit; the integrated units may be implemented in hardware or in hardware plus software functional units.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, where the program, when executed, performs steps including the above method embodiments; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk or an optical disk, or the like, which can store program codes.
Alternatively, the above-described integrated units of the present invention may be stored in a computer-readable storage medium if implemented in the form of software functional modules and sold or used as separate products. Based on such understanding, the technical solutions of the embodiments of the present invention may be embodied in essence or a part contributing to the prior art in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, ROM, RAM, magnetic or optical disk, or other medium capable of storing program code.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (9)
1. The diffusion prediction method based on deep learning is characterized by comprising the following steps of:
acquiring pollution source information in a target area, screening target pollutants according to the pollution source information, and determining target pollutant monitoring points according to GIS geographic data of the target area;
collecting target pollutant concentration information through target pollutant monitoring points, and clustering all monitoring points based on the target pollutant concentration information to obtain a pollution diffusion pattern of a target area;
acquiring a concentration change sequence of target pollutants in a pollution area according to a pollution diffusion pattern of the target area, constructing a diffusion prediction model, and training based on the concentration change sequence of the target pollutants in the pollution area;
acquiring space-time correlation characteristics of the concentration of the target pollutant, the soil property and the water quality property in a continuous time period through a diffusion prediction model, and representing the diffusion trend of the target pollutant through the space-time correlation characteristics;
The potential pollution area is obtained by combining the pollution distribution of the current time of the target area with the diffusion trend of the target pollutant, and the predicted pollution diffusion distribution is generated and visually displayed;
screening target pollutants according to the pollution source information, and determining target pollutant monitoring points according to GIS geographic data of a target area, wherein the target pollutant monitoring points are specifically as follows:
acquiring pollution source information in a range according to the regional range of the target region, extracting pollution source position information and pollutants produced by a pollution source according to the pollution source information, and screening out marked pollutants of the target region through the pollutants;
screening in the marked pollutants according to a preset environment monitoring task to obtain target pollutants, obtaining position information of an existing environment monitoring well of a target area, extracting environment monitoring information in the environment monitoring well, and analyzing the existing monitoring data of the target pollutants;
carrying out data preprocessing on the existing monitoring data of the target pollutants, carrying out data analysis on the preprocessed existing monitoring data, retrieving and acquiring experience related to the diffusion of the target pollutants by a big data means, and setting a monitoring data correlation standard under the current environmental monitoring well position distribution;
acquiring data association of environmental monitoring wells at different positions according to data analysis, judging the deviation of the data association and a monitoring data correlation standard, and acquiring GIS geographic data of a target area when the deviation is not in a preset deviation interval;
And acquiring the terrain and hydrologic data in the target area according to the GIS geographic data, carrying out position optimization and newly adding monitoring points by utilizing a genetic algorithm on the basis of the existing environment monitoring well according to the pollution source position information, and generating target pollutant monitoring points by combining the environment monitoring well.
2. The diffusion prediction method based on deep learning according to claim 1, wherein the target pollutant concentration information is collected through target pollutant monitoring points, and clustering is performed on each monitoring point based on the target pollutant concentration information, so as to obtain a pollution diffusion pattern of a target area, which specifically comprises:
dividing a target area into grids according to the position information of target pollutant monitoring points, acquiring target pollutant concentration information acquired by the target pollutant monitoring points, and acquiring a clustering sample set by combining the grids with the target pollutant concentration information;
obtaining a similarity matrix between grid samples according to grid position information and target pollutant concentration information in a cluster sample set, obtaining attraction and attribution between the grid samples through the similarity matrix, and obtaining a cluster center of the cluster sample set after multiple iterations;
obtaining a clustering result through a clustering center of the clustering sample set, obtaining pollution distribution of a target area according to the clustering result, and determining a diffusion direction according to intra-class distances of all grid samples in the clustering result;
And generating a diffusion pattern of the target pollutant in the target area according to the pollution distribution and the diffusion direction of the target area.
3. The diffusion prediction method based on deep learning according to claim 1, wherein the time-space correlation characteristics of the target pollutant concentration, the soil property and the water quality property in the continuous time period are obtained by using a diffusion prediction model, specifically:
acquiring a concentration change sequence of the target pollutant in the polluted area according to the historical monitoring data of the target pollutant and the diffusion pattern of the target pollutant, preprocessing the concentration change sequence, and removing abnormal monitoring data;
constructing a diffusion prediction model based on deep learning, taking the pretreated target pollutant concentration change sequence as a model input, and carrying out normalization treatment;
learning and representing a graph structure formed by target pollutant monitoring points in a target area through a graph attention network, constructing an adjacent matrix to acquire a spatial relationship between adjacent target pollutant monitoring points, and distributing different weights to neighbor nodes by combining average pooling and maximum pooling;
in addition, attention coefficients are set among target pollutant monitoring stations according to the diffusion directions, aggregation of neighbor nodes is carried out according to the attention coefficients, useless nodes are filtered, and spatial features corresponding to the target pollutant concentration change sequences are obtained according to a neighbor aggregation mechanism of a graph attention network;
Inputting a characteristic sequence output by a graph attention network into a DASEQ2Seq model for characteristic coding, acquiring the hidden state of the current timestamp by utilizing the characteristic input of the current timestamp and the hidden state of the previous timestamp, introducing time attention into the DASEQ2Seq model, and representing the importance of each characteristic in the characteristic sequence;
and after decoding by a decoder, acquiring the space-time correlation characteristic.
4. The diffusion prediction method based on deep learning as claimed in claim 3, wherein the graph structure formed by the target pollutant monitoring points of the target area is subjected to learning representation through a graph attention network, specifically: acquiring position information of target pollutant monitoring points in a target area, and determining a spatial topological relationDefined as->，/>Representing target contaminant monitoring points as nodes in the graph structure, +.>Representing a spatial relationship between target contaminant monitoring points; acquiring a grid where target pollutant monitoring points are located, taking corresponding target pollutant monitoring data as node characteristics, and acquiring soil characteristics and water characteristics in the grid as additional characteristics of the nodes;
and performing representation learning on the node characteristics through a message passing mechanism and a neighbor aggregation mechanism of the graph attention network, and generating a corresponding low-dimensional vector representation.
5. The diffusion prediction method based on deep learning according to claim 1, wherein the potential pollution area is obtained by combining the pollution distribution of the current time of the target area with the diffusion trend of the target pollutant, and the predicted pollution diffusion distribution is generated and visually displayed, specifically:
acquiring current pollution distribution of the target pollutant according to a pollution diffusion pattern in the target area, and inputting the current pollution distribution into a trained diffusion prediction model;
extracting space-time correlation characteristics through the diffusion prediction model, and acquiring target pollutant prediction concentrations of target pollutant monitoring points after preset time according to the space-time correlation characteristics;
performing cluster analysis according to the target pollutant predicted concentration to generate a new cluster result, and comparing the new cluster result with a historical cluster result to obtain a cluster deviation;
and determining a corresponding grid according to the clustering deviation, marking, generating predicted pollution diffusion distribution by using the marked grid potential pollution area, generating related early warning information according to the predicted pollution diffusion distribution, and carrying out visual processing on the early warning information and the predicted pollution diffusion distribution.
6. A diffusion prediction system based on deep learning, the system comprising: the diffusion prediction method based on the deep learning comprises a memory and a processor, wherein the memory comprises a diffusion prediction method program based on the deep learning, and the diffusion prediction method program based on the deep learning realizes the following steps when being executed by the processor:
acquiring pollution source information in a target area, screening target pollutants according to the pollution source information, and determining target pollutant monitoring points according to GIS geographic data of the target area;
collecting target pollutant concentration information through target pollutant monitoring points, and clustering all monitoring points based on the target pollutant concentration information to obtain a pollution diffusion pattern of a target area;
acquiring a concentration change sequence of target pollutants in a pollution area according to a pollution diffusion pattern of the target area, constructing a diffusion prediction model, and training based on the concentration change sequence of the target pollutants in the pollution area;
acquiring space-time correlation characteristics of the concentration of the target pollutant, the soil property and the water quality property in a continuous time period through a diffusion prediction model, and representing the diffusion trend of the target pollutant through the space-time correlation characteristics;
the potential pollution area is obtained by combining the pollution distribution of the current time of the target area with the diffusion trend of the target pollutant, and the predicted pollution diffusion distribution is generated and visually displayed;
Screening target pollutants according to the pollution source information, and determining target pollutant monitoring points according to GIS geographic data of a target area, wherein the target pollutant monitoring points are specifically as follows:
acquiring pollution source information in a range according to the regional range of the target region, extracting pollution source position information and pollutants produced by a pollution source according to the pollution source information, and screening out marked pollutants of the target region through the pollutants;
screening in the marked pollutants according to a preset environment monitoring task to obtain target pollutants, obtaining position information of an existing environment monitoring well of a target area, extracting environment monitoring information in the environment monitoring well, and analyzing the existing monitoring data of the target pollutants;
carrying out data preprocessing on the existing monitoring data of the target pollutants, carrying out data analysis on the preprocessed existing monitoring data, retrieving and acquiring experience related to the diffusion of the target pollutants by a big data means, and setting a monitoring data correlation standard under the current environmental monitoring well position distribution;
acquiring data association of environmental monitoring wells at different positions according to data analysis, judging the deviation of the data association and a monitoring data correlation standard, and acquiring GIS geographic data of a target area when the deviation is not in a preset deviation interval;
And acquiring the terrain and hydrologic data in the target area according to the GIS geographic data, carrying out position optimization and newly adding monitoring points by utilizing a genetic algorithm on the basis of the existing environment monitoring well according to the pollution source position information, and generating target pollutant monitoring points by combining the environment monitoring well.
7. The diffusion prediction system based on deep learning according to claim 6, wherein the target pollutant concentration information is collected by the target pollutant monitoring points, and the pollution distribution of the target area is obtained by clustering each monitoring point based on the target pollutant concentration information, which is specifically as follows:
dividing a target area into grids according to the position information of target pollutant monitoring points, acquiring target pollutant concentration information acquired by the target pollutant monitoring points, and acquiring a clustering sample set by combining the grids with the target pollutant concentration information;
obtaining a similarity matrix between grid samples according to grid position information and target pollutant concentration information in a cluster sample set, obtaining attraction and attribution between the grid samples through the similarity matrix, and obtaining a cluster center of the cluster sample set after multiple iterations;
obtaining a clustering result through a clustering center of the clustering sample set, obtaining pollution distribution of a target area according to the clustering result, and determining a diffusion direction according to intra-class distances of all grid samples in the clustering result;
And generating a diffusion pattern of the target pollutant in the target area according to the pollution distribution and the diffusion direction of the target area.
8. The diffusion prediction system based on deep learning according to claim 6, wherein the time-space correlation features are obtained for the concentration of the target pollutant and the soil property and the water quality property in the continuous time period through the diffusion prediction model, specifically:
acquiring a concentration change sequence of the target pollutant in the polluted area according to the historical monitoring data of the target pollutant and the diffusion pattern of the target pollutant, preprocessing the concentration change sequence, and removing abnormal monitoring data;
constructing a diffusion prediction model based on deep learning, taking the pretreated target pollutant concentration change sequence as a model input, and carrying out normalization treatment;
learning and representing a graph structure formed by target pollutant monitoring points in a target area through a graph attention network, constructing an adjacent matrix to acquire a spatial relationship between adjacent target pollutant monitoring points, and distributing different weights to neighbor nodes by combining average pooling and maximum pooling;
in addition, attention coefficients are set among target pollutant monitoring stations according to the diffusion directions, aggregation of neighbor nodes is carried out according to the attention coefficients, useless nodes are filtered, and spatial features corresponding to the target pollutant concentration change sequences are obtained according to a neighbor aggregation mechanism of a graph attention network;
Inputting a characteristic sequence output by a graph attention network into a DASEQ2Seq model for characteristic coding, acquiring the hidden state of the current timestamp by utilizing the characteristic input of the current timestamp and the hidden state of the previous timestamp, introducing time attention into the DASEQ2Seq model, and representing the importance of each characteristic in the characteristic sequence;
and after decoding by a decoder, acquiring the space-time correlation characteristic.
9. The diffusion prediction system based on deep learning according to claim 6, wherein the potential pollution area is obtained by combining the pollution distribution of the current time of the target area with the diffusion trend of the target pollutant, and the predicted pollution diffusion distribution is generated and visually displayed, specifically:
acquiring current pollution distribution of the target pollutant according to a pollution diffusion pattern in the target area, and inputting the current pollution distribution into a trained diffusion prediction model;
extracting space-time correlation characteristics through the diffusion prediction model, and acquiring target pollutant prediction concentrations of target pollutant monitoring points after preset time according to the space-time correlation characteristics;
performing cluster analysis according to the target pollutant predicted concentration to generate a new cluster result, and comparing the new cluster result with a historical cluster result to obtain a cluster deviation;
And determining a corresponding grid according to the clustering deviation, marking, generating predicted pollution diffusion distribution by using the marked grid potential pollution area, generating related early warning information according to the predicted pollution diffusion distribution, and carrying out visual processing on the early warning information and the predicted pollution diffusion distribution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310465760.9A CN116186566B (en) | 2023-04-27 | 2023-04-27 | Diffusion prediction method and system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310465760.9A CN116186566B (en) | 2023-04-27 | 2023-04-27 | Diffusion prediction method and system based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116186566A CN116186566A (en) | 2023-05-30 |
| CN116186566B true CN116186566B (en) | 2023-07-21 |
Family
ID=86434863
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310465760.9A Active CN116186566B (en) | 2023-04-27 | 2023-04-27 | Diffusion prediction method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116186566B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116384754B (en) * | 2023-06-02 | 2023-08-04 | 北京建工环境修复股份有限公司 | Deep learning-based environmental pollution risk assessment method in chemical industry park |
| CN116362569B (en) * | 2023-06-02 | 2023-08-01 | 北京建工环境修复股份有限公司 | Multi-dimensional evaluation method and system for site pollution |
| CN116362571B (en) * | 2023-06-02 | 2023-08-08 | 北京建工环境修复股份有限公司 | Multi-source pollution space-time characteristic analysis method and system based on deep learning |
| CN116384775B (en) * | 2023-06-02 | 2023-08-01 | 北京建工环境修复股份有限公司 | Management method, system and storage medium of pollution site monitoring equipment |
| CN116384776B (en) * | 2023-06-02 | 2023-08-01 | 北京建工环境修复股份有限公司 | Site pollution evaluation method, system and storage medium based on fuzzy evaluation |
| CN116629035B (en) * | 2023-07-25 | 2023-09-15 | 欧梯恩智能科技(苏州)有限公司 | River pollution diffusion simulation method, device and system |
| CN116773781B (en) * | 2023-08-18 | 2023-12-05 | 北京建工环境修复股份有限公司 | Pollution analysis method, system and medium for perfluorinated compounds in soil |
| CN116842350B (en) * | 2023-09-01 | 2023-11-03 | 北京建工环境修复股份有限公司 | Analysis method, system and medium for phytoremediation of polluted site |
| CN116975378B (en) * | 2023-09-21 | 2023-12-05 | 江苏云天精筑环境科技有限公司 | Equipment environment monitoring method and system based on big data |
| CN117390464B (en) * | 2023-12-11 | 2024-03-08 | 唐山华洋自动化有限公司 | Distributed atmosphere pollution monitoring method and system based on KAFKA technology |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107145611A (en) * | 2017-06-15 | 2017-09-08 | 阿克苏地区环境保护局 | Priority industry pollution sources monitor dynamic management information platform in real time |
| CN113657698A (en) * | 2020-05-12 | 2021-11-16 | 中国环境科学研究院 | Basin partition pollution source identification method based on multivariate statistics and receptor model |
| US20230007082A1 (en) * | 2021-06-30 | 2023-01-05 | Microsoft Technology Licensing, Llc | Pollutant sensor placement |
-
2023
- 2023-04-27 CN CN202310465760.9A patent/CN116186566B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN116186566A (en) | 2023-05-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116186566B (en) | Diffusion prediction method and system based on deep learning | |
| Kang et al. | Air quality prediction: Big data and machine learning approaches | |
| CN115578015B (en) | Sewage treatment whole process supervision method, system and storage medium based on Internet of things | |
| CN116385689B (en) | Visual information management method, system and medium for site pollution data | |
| CN108009674A (en) | Air PM2.5 concentration prediction methods based on CNN and LSTM fused neural networks | |
| CN116629619B (en) | Method and system for predicting occurrence condition of new soil pollutants | |
| CN116384754B (en) | Deep learning-based environmental pollution risk assessment method in chemical industry park | |
| CN116485202B (en) | Industrial pollution real-time monitoring method and system based on Internet of things | |
| CN116362569B (en) | Multi-dimensional evaluation method and system for site pollution | |
| CN117171223B (en) | Microorganism culture scheme recommendation method and system in microorganism repair process | |
| CN115308101B (en) | Water quality monitoring management method and system | |
| CN116168771B (en) | Method and system for monitoring pollution of perfluorinated compounds in river based on Internet of things | |
| CN113033861A (en) | Water quality prediction method and system based on time series model | |
| CN115017511A (en) | Source code vulnerability detection method and device and storage medium | |
| Luo et al. | A fast PM2. 5 forecast approach based on time-series data analysis, regression and regularization | |
| CN116384775B (en) | Management method, system and storage medium of pollution site monitoring equipment | |
| CN116090710B (en) | Management method, system, electronic equipment and medium for enterprise pollution discharge permission | |
| CN116233865B (en) | Point distribution method and system for new pollutant monitoring equipment | |
| CN117033923A (en) | Method and system for predicting crime quantity based on interpretable machine learning | |
| CN115963243A (en) | Soil nitrogen and phosphorus nutrient loss prevention and control method and system based on nutrient dynamic monitoring | |
| Mohammad et al. | Air pollution comparison RFM model using machine learning approach | |
| CN117172996B (en) | Microbial activity identification monitoring method and system for ecological environment restoration | |
| CN117172578B (en) | Soil ecological environment restoration method and system based on microorganism identification | |
| CN117171695B (en) | Method and system for evaluating ecological restoration effect of antibiotic contaminated soil | |
| CN117542434B (en) | Air pollutant concentration analysis method based on data analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |