Historical landscape risk assessment and early warning system based on big data

Abstract

The invention discloses a historical landscape risk assessment and early warning system based on big data, relates to the field of data mining and risk management, and solves the problems that feature extraction is difficult to carry out on various preprocessed data, corresponding pollution risk indexes are difficult to obtain through model construction and calculation, and the risk assessment and early warning efficiency is high. In order to solve the technical problems that a historical landscape risk prediction model is difficult to construct through various characteristic pollution risk indexes, and classification and early warning are performed on historical landscape risk degrees, the system comprises the following modules: a data acquisition and processing module, which is used for acquiring and preprocessing various data in combination with a remote sensing GIS (Geographic Information System) technology; the data analysis module is used for analyzing data to obtain an air pollution risk index, a terrain risk index and a comprehensive biological imbalance risk index; the feature model prediction module is used for constructing a historical landscape risk prediction model by using multiple risk indexes to obtain a prediction value; and the risk assessment and early warning module is used for carrying out classification and early warning on the risk degrees of the historical landscapes.

Description

Technical Field

[0001] The present invention belongs to the field of data mining and risk management, and specifically relates to a historical landscape risk assessment and early warning system based on big data. Background Art

[0002] As the impact of global climate change on the environment becomes increasingly severe, the protection of historical landscapes faces unprecedented challenges. In order to effectively assess these risks, a historical landscape risk assessment and early warning system based on big data is provided. The corresponding risk index is obtained by extracting characteristic data from a variety of collected data, including air pollution risk index, terrain risk index and comprehensive biological imbalance risk index. The corresponding risk index is used as input data to construct a historical landscape risk prediction model, which is deployed in the historical landscape environment to obtain a comprehensive historical landscape risk assessment prediction value. Different risk thresholds are set, and the risk level of the historical landscape is classified and warned according to the comprehensive historical landscape risk assessment prediction value.

[0003] Although the existing historical landscape risk assessment and early warning system has achieved historical landscape risk assessment and early warning to a certain extent, it is difficult for the traditional historical landscape risk assessment and early warning system to analyze a variety of historical data through big data analysis methods. It lacks the extraction of corresponding features from the collected air environment data, topographic data and biological sample data to obtain the air pollution risk index, topographic risk index and comprehensive biological imbalance risk index respectively. It is also difficult to integrate multiple characteristic pollution risk indices to construct a historical landscape risk prediction model to obtain a comprehensive historical landscape risk evaluation prediction value. It also lacks classification and early warning of the risk level of historical landscapes. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a historical landscape risk assessment and early warning system based on big data, which is used to solve the following technical problems: Although the existing historical landscape risk assessment and early warning system has achieved historical landscape risk assessment and early warning to a certain extent, it is difficult for the traditional historical landscape risk assessment and early warning system to analyze a variety of historical data through big data analysis methods. It lacks the extraction of corresponding features from the collected air environment data, topographic data and biological sample data to obtain the air pollution risk index, topographic risk index and comprehensive biological imbalance risk index respectively. It is also difficult to integrate multiple characteristic pollution risk indices to construct a historical landscape risk prediction model to obtain a comprehensive historical landscape risk evaluation prediction value. It also lacks classification and early warning of the risk level of historical landscapes.

[0005] To solve the above problems, the first aspect of the present invention provides a historical landscape risk assessment and early warning system based on big data, comprising the following modules: Data collection and processing module: collects a variety of data through the combination of remote sensing technology and GIS technology, including air environment monitoring data, topographic data and biological sample data, and pre-processes the data; Data analysis module: The relative offset values ​​of the corresponding pollutants obtained by comparing the measured values ​​of pollutants with the average values ​​of pollutant concentrations are used to obtain the air pollution risk index. The rainfall prediction model of the historical landscape is constructed through the historical rainfall data. The terrain complexity is obtained by comparing the rainfall prediction value with the terrain complexity to obtain the terrain risk index. The biological sample data obtained from the historical landscape are used to construct a biological recognition model. The comprehensive biological imbalance risk index is obtained based on the plant imbalance index, animal imbalance index and microbial imbalance index. Characteristic model prediction module: The characteristic data extracted from the collected air environment monitoring data, topographic data and biological sample data are used to obtain the air pollution risk index, topographic risk index and comprehensive biological imbalance risk index to build a model to obtain the comprehensive historical landscape risk assessment prediction value; Risk assessment and early warning module: Set different risk thresholds, and classify and warn the risk levels of historical landscapes based on the predicted values ​​of comprehensive historical landscape risk assessment.

[0006] As a further solution of the present invention, the method of collecting a variety of data, including air environment monitoring data, topographic data and biological sample data, by combining remote sensing technology and GIS technology, and preprocessing the data includes the following steps: After using satellite sensors to obtain satellite image data, GPS technology is used to measure the position of ground control points, and GPS receivers are used to measure the coordinates of each ground control point. LiDAR technology is used to obtain LiDAR data to collect topographic data and process them to obtain a DEM digital elevation model. Satellite sensors are used to monitor the types of pollutants, and air quality detection instruments are deployed in historical landscapes to collect real-time data. Plant, animal and biological samples are collected in historical landscape environments. High-resolution images are used to identify and classify different biological species, and the collected data is preprocessed, including geometric correction and denoising.

[0007] As a further solution of the present invention, the relative deviation value of the corresponding pollutant obtained by comparing the measured value of the pollutant with the average value of the pollutant concentration, and obtaining the air pollution risk index through the relative deviation values ​​of multiple pollutants, include the following steps: By deploying multiple air quality monitoring stations in the historical landscape, each station is equipped with a particle detector and a gaseous pollutant detector to collect data on PM2.5, SO2, NO2 and CO pollutants respectively, regularly monitor and record the corresponding concentration data of pollutants in the air, and after pre-processing the collected data, use big data analysis to obtain the average concentration of each pollutant. The relative offset value of the corresponding pollutant obtained by the pollutant measurement value collected by the measuring instrument and the average pollutant concentration is used to obtain the air pollution risk index through the relative offset values ​​of multiple pollutants; Air pollution risk index calculation formula: in, is the air pollution risk index, is the measured value of PM2.5 concentration in the air, is the average value of PM2.5 concentration, is the measured value of SO2 concentration in the air, is the average value of SO2 concentration, is the measured value of NO2 concentration in the air, is the average value of NO2 concentration, is the measured value of CO concentration in the air, is the average value of CO concentration.

[0008] As a further solution of the present invention, the rainfall prediction model of the historical landscape is constructed by using the historical rainfall data, the terrain complexity is obtained by the feature data extracted from the remote sensing image data, and the terrain risk index is obtained by the rainfall prediction value and the terrain complexity, including the following steps: The remote sensing image data of the historical landscape is obtained through remote sensing technology, and the historical rainfall data of the historical landscape is obtained from the meteorological department. The time series features are extracted from the historical rainfall data using statistical feature analysis methods, including annual rainfall, average monthly rainfall and maximum daily rainfall. The extracted feature data is used as input variables, and the historical landscape rainfall data is used as output variables to train the model. The new remote sensing image data and feature data are input into the trained model. The model outputs the rainfall prediction value of the historical landscape based on the input data. The acquired remote sensing image data was preprocessed, including noise reduction, atmospheric correction and image enhancement. The terrain features, including slope, aspect and height features, were extracted from the preprocessed data. The slope tool in ArcGIS geographic information system was used to calculate the elevation change between each pixel and the surrounding 3×3 pixel neighborhood based on the DEM elevation model data. The corresponding elevation change rate was determined by comparing the elevation values ​​of the central pixel and the adjacent pixels to obtain the slope value. The aspect tool in ArcGIS geographic information system was used to calculate the value of each pixel based on the DEM elevation model data to represent the direction of the point. The aspect was classified by the reclassification tool, and a numerical value was set for each category to represent it. The ground height information was extracted from the DEM elevation model data, and the elevation value of each pixel was obtained and input into the attribute table. The height value was extracted from the attribute table using the raster calculator. The slope value, aspect value and height value were standardized and then combined to obtain the terrain complexity. The calculation formula of terrain complexity is: in, is the complexity of the terrain, is the slope value, is the slope value, is the height value; The predicted rainfall values ​​of historical landscapes and the terrain complexity were normalized and integrated to obtain the terrain risk index; Terrain risk index calculation formula: in, is the terrain risk index, is the rainfall forecast value, and are all weight coefficients.

[0009] As a further solution of the present invention, the biometric recognition model is constructed by using the biological sample data obtained from the historical landscape, comprising the following steps: The biological sample data obtained from the historical landscape include the number and types of all species in the historical landscape. The plant, animal and microbial samples in the biological sample data are labeled and classified respectively. The tidyverse neat data science package in the R language is used to clean the data, including processing missing values, outliers and normalizing the data. The corresponding features of different types of biological samples are extracted through image processing software and text analysis methods. For plant samples, leaf shape, color and texture features are extracted. For animal samples, animal body shape, color and habitat features are extracted. For microbial samples, cell shape, size and color features are extracted. The statistical functions in the Python programming language are used to calculate the number of each type of biological samples. The corresponding biometric recognition model is constructed according to the features extracted from the data, and the labeled data is used to train the model. The features extracted from the biological samples are used as input, and the types and corresponding quantities of biological samples are used as output. The trained model is deployed in a new environment to identify new biological samples.

[0010] As a further solution of the present invention, obtaining a comprehensive biological imbalance risk index according to the plant imbalance index, the animal imbalance index and the microbial imbalance index comprises the following steps: The species and quantity data of various biological samples obtained through the model output were used to obtain the species uniformity decline index and species richness decline index of plants, animals and microorganisms respectively. The species imbalance index corresponding to different biological samples was obtained according to the average values ​​of the species uniformity decline index and species richness decline index of different biological samples respectively. The comprehensive biological imbalance risk index was obtained according to the plant imbalance index, animal imbalance index and microbial imbalance index. Species imbalance index calculation formula: in, is the species imbalance index, When is the plant imbalance index; When is the animal imbalance index; When is the microbial imbalance index; Any of plants, animals and microorganisms The number of individuals of a species, is the total number of individuals of any species of plants, animals and microorganisms, The total number of any one type of plant, animal and microorganism; Calculation formula for comprehensive biological imbalance risk index: in, is the comprehensive biological imbalance risk index, is the plant imbalance index, is the animal imbalance index, is the microbial imbalance index, , and are the corresponding weight coefficients.

[0011] As a further solution of the present invention, the air pollution risk index, terrain risk index and comprehensive biological imbalance risk index are respectively obtained by extracting characteristic data from the collected air environment monitoring data, topographic data and biological sample data to construct a model to obtain a comprehensive historical landscape risk assessment prediction value, including the following steps: The air pollution risk index, terrain risk index and comprehensive biological imbalance risk index are obtained by extracting characteristic data from the collected air environment monitoring data, topographic data and biological sample data. These data are used as the input features of the model, and the comprehensive historical landscape risk assessment prediction value is used as the output data to construct a historical landscape risk prediction model. The model is trained using the historical air pollution risk index, terrain risk index and comprehensive biological imbalance risk index. The trained model is deployed to the new historical landscape for real-time risk analysis.

[0012] As a further solution of the present invention, the setting of different risk thresholds and the classification and early warning of the risk levels of historical landscapes according to the predicted values ​​of the comprehensive historical landscape risk assessment include the following steps: Different risk thresholds are set according to the distribution of historical data; if 0<comprehensive historical landscape risk assessment prediction value<0.2, it indicates a low risk level and a third-level warning signal is issued; if 0.2≤comprehensive historical landscape risk assessment prediction value<0.5, it indicates a medium risk level and a second-level warning signal is issued; if 0.5≤comprehensive historical landscape risk assessment prediction value, it indicates a high risk level and a first-level warning signal is issued.

[0013] Beneficial effects of the present invention: The present invention obtains the average concentration of each pollutant through a big data analysis method, obtains the relative offset value of the corresponding pollutant by comparing the pollutant measurement value collected by the measuring instrument with the average pollutant concentration value, and obtains a comprehensive air pollution risk index through the relative offset values ​​of multiple pollutants. The statistical feature analysis method is used to extract time series features and remote sensing image data from historical rainfall data as input variables to build a historical landscape rainfall prediction model, and the terrain risk index is obtained after normalizing the rainfall prediction value with the terrain features extracted from the remote sensing image data. The plant, animal and microbial samples in the biological sample data are respectively labeled and classified to extract the corresponding feature data, and the statistical function in the Python programming language is used to calculate the number of each type of biological sample, and a sample recognition model is constructed according to the number and feature data to identify new biological samples, and the types and quantities of various biological samples obtained through the model output are obtained. The data were used to obtain the species evenness decline index and species richness decline index of plants, animals and microorganisms respectively. The species imbalance index of different biological samples was obtained according to the average values ​​of the species evenness decline index and species richness decline index of different biological samples, and the comprehensive biological imbalance risk index was obtained. A historical landscape risk prediction model was constructed by using the air pollution risk index, terrain risk index and comprehensive biological imbalance risk index as characteristic data. The trained model was deployed to the new historical landscape to obtain the comprehensive historical landscape risk assessment prediction value, and different risk thresholds were set up. According to the comprehensive historical landscape risk assessment prediction value, the risk level of the historical landscape was classified and warned, which realized the integration of multiple risk factors and the comprehensive assessment of the comprehensive risk of the historical landscape. This multi-dimensional assessment method can more accurately reflect the actual risk status of the historical landscape. By setting different risk thresholds, dynamic monitoring and early warning of historical landscape risks can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 work.

[0015] Figure 1 It is a schematic diagram of the system module of the present invention. DETAILED DESCRIPTION

[0016] 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.

[0017] See also Figure 1 As shown, the present invention is a historical landscape risk assessment and early warning system based on big data, which includes the following modules: data acquisition and processing module: a variety of data are collected through the combination of remote sensing technology and GIS technology, including air environment monitoring data, topographic data and biological sample data, and the data are pre-processed; Data analysis module: The relative offset values ​​of the corresponding pollutants obtained by comparing the measured values ​​of pollutants with the average values ​​of pollutant concentrations are used to obtain the air pollution risk index. The rainfall prediction model of the historical landscape is constructed through the historical rainfall data. The terrain complexity is obtained by comparing the rainfall prediction value with the terrain complexity to obtain the terrain risk index. The biological sample data obtained from the historical landscape are used to construct a biological recognition model. The comprehensive biological imbalance risk index is obtained based on the plant imbalance index, animal imbalance index and microbial imbalance index. Characteristic model prediction module: The characteristic data extracted from the collected air environment monitoring data, topographic data and biological sample data are used to obtain the air pollution risk index, topographic risk index and comprehensive biological imbalance risk index to build a model to obtain the comprehensive historical landscape risk assessment prediction value; Risk assessment and early warning module: Set different risk thresholds, and classify and warn the risk levels of historical landscapes based on the predicted values ​​of comprehensive historical landscape risk assessment.

[0018] Specifically, a variety of data are collected by combining multiple technologies and preprocessed. The relative offset value of the corresponding pollutant is obtained by the ratio of the pollutant measurement value to the average pollutant concentration, and finally the air pollution risk index is obtained. The time series feature data extracted from the historical rainfall data is used as input and the historical landscape rainfall is used as output to construct a rainfall prediction model for the historical landscape to obtain the rainfall prediction value. The terrain complexity obtained by the feature data extracted from the remote sensing image data and the rainfall prediction value of the historical landscape are normalized and integrated to obtain the terrain risk index. The biological sample data obtained from the historical landscape are annotated and classified respectively for the plant, animal and microbial samples in the biological sample data, and the corresponding feature data are extracted to construct a biometric recognition model. The comprehensive biological imbalance risk index is obtained according to the plant, animal and microbial imbalance index. The model is constructed by three pollution risk indices to obtain the comprehensive historical landscape risk assessment prediction value. Different risk thresholds are set to classify and warn the risk level of the historical landscape.

[0019] In one embodiment of the present invention, the method of collecting a variety of data, including air environment monitoring data, topographic data and biological sample data, by combining remote sensing technology and GIS technology, and preprocessing the data includes the following steps: After using satellite sensors to obtain satellite image data, GPS technology is used to measure the position of ground control points, and GPS receivers are used to measure the coordinates of each ground control point. LiDAR technology is used to obtain LiDAR data to collect topographic data and process them to obtain a DEM digital elevation model. Satellite sensors are used to monitor the types of pollutants, and air quality detection instruments are deployed in historical landscapes to collect real-time data. Plant, animal and biological samples are collected in historical landscape environments. High-resolution images are used to identify and classify different biological species, and the collected data is preprocessed, including geometric correction and denoising.

[0020] Specifically, after using satellite sensors to obtain satellite image data, use GPS technology to measure the positions of ground control points to ensure that they are evenly distributed and sufficient in number, use a GPS receiver to measure the coordinates of each ground control point, record the longitude, latitude and altitude information, obtain lidar data through lidar technology, collect terrain data and process them to obtain a DEM digital elevation model, use satellite sensors with environmental monitoring capabilities to detect the types of pollutants, analyze pollutant concentration data, collect plant, animal and microbial sample data in historical landscapes, and record relevant information, use remote sensing software to pre-process images, use the GCP cloud computing service platform to perform geometric correction on satellite images to eliminate geometric distortion caused by the earth's curvature and sensor attitude factors, and apply filters to denoise the images to improve image quality.

[0021] In one embodiment of the present invention, the relative offset value of the corresponding pollutant obtained by comparing the pollutant measurement value with the average value of the pollutant concentration, and obtaining the air pollution risk index through the relative offset values ​​of multiple pollutants, comprises the following steps: By deploying multiple air quality monitoring stations in the historical landscape, each station is equipped with a particle detector and a gaseous pollutant detector to collect data on PM2.5, SO2, NO2 and CO pollutants respectively, regularly monitor and record the corresponding concentration data of pollutants in the air, and after pre-processing the collected data, use big data analysis to obtain the average concentration of each pollutant. The relative offset value of the corresponding pollutant obtained by the pollutant measurement value collected by the measuring instrument and the average pollutant concentration is used to obtain the air pollution risk index through the relative offset values ​​of multiple pollutants; Air pollution risk index calculation formula: in, is the air pollution risk index, is the measured value of PM2.5 concentration in the air, is the average value of PM2.5 concentration, is the measured value of SO2 concentration in the air, is the average value of SO2 concentration, is the measured value of NO2 concentration in the air, is the average value of NO2 concentration, is the measured value of CO concentration in the air, is the average value of CO concentration.

[0022] Specifically, by deploying multiple air quality monitoring stations in the historical landscape, each station is equipped with a particle detector and a gaseous pollutant detector. The particle detector is used to detect PM2.5, and the gaseous pollutant detector is used to detect SO2, NO2 and CO. The data of each detection station is regularly monitored and collected, and the corresponding concentration data of pollutants in the air are recorded. After removing invalid data points, the collected data are normalized and all data are converted into a unified dimension for subsequent analysis. For each pollutant, the corresponding concentration average is calculated using big data analysis. The actual pollutant concentration value collected by the measuring instrument is compared with its corresponding average value, and the relative offset is calculated, which reflects the degree to which the actual concentration deviates from the average level. According to the relative offset value of each pollutant and its weight, the air pollution risk index is calculated to assess the current air quality status and potential landscape air pollution risks.

[0023] In one embodiment of the present invention, the construction of a rainfall prediction model for a historical landscape through historical rainfall data, obtaining a terrain complexity from feature data extracted from remote sensing image data, and obtaining a terrain risk index through rainfall prediction values ​​and terrain complexity include the following steps: The remote sensing image data of the historical landscape is obtained through remote sensing technology, and the historical rainfall data of the historical landscape is obtained from the meteorological department. The time series features are extracted from the historical rainfall data using statistical feature analysis methods, including annual rainfall, average monthly rainfall and maximum daily rainfall. The extracted feature data is used as input variables, and the historical landscape rainfall data is used as output variables to train the model. The new remote sensing image data and feature data are input into the trained model. The model outputs the rainfall prediction value of the historical landscape based on the input data. The acquired remote sensing image data was preprocessed, including noise reduction, atmospheric correction and image enhancement. The terrain features, including slope, aspect and height features, were extracted from the preprocessed data. The slope tool in ArcGIS geographic information system was used to calculate the elevation change between each pixel and the surrounding 3×3 pixel neighborhood based on the DEM elevation model data. The corresponding elevation change rate was determined by comparing the elevation values ​​of the central pixel and the adjacent pixels to obtain the slope value. The aspect tool in ArcGIS geographic information system was used to calculate the value of each pixel based on the DEM elevation model data to represent the direction of the point. The aspect was classified by the reclassification tool, and a numerical value was set for each category to represent it. The ground height information was extracted from the DEM elevation model data, and the elevation value of each pixel was obtained and input into the attribute table. The height value was extracted from the attribute table using the raster calculator. The slope value, aspect value and height value were standardized and then combined to obtain the terrain complexity. The calculation formula of terrain complexity is: in, is the complexity of the terrain, is the slope value, is the slope value, is the height value; The predicted rainfall values ​​of historical landscapes and the terrain complexity were normalized and integrated to obtain the terrain risk index; Terrain risk index calculation formula: in, is the terrain risk index, is the rainfall forecast value, and are all weight coefficients.

[0024] Specifically, the remote sensing tool in the ArcGIS geographic information system is used to import and process the remote sensing image data of the historical landscape, and the historical rainfall data of the historical landscape is obtained from the meteorological department, including annual rainfall, average monthly rainfall and maximum daily rainfall data. The big data analysis method is used to extract time series characteristics from the historical rainfall data of the historical landscape, and the statistical indicators of annual rainfall, average monthly rainfall and maximum daily rainfall are calculated. The extracted time series characteristic data and remote sensing image data are used as input variables, and the historical landscape rainfall data is used as the output variable to construct a historical landscape rainfall prediction model to obtain the rainfall prediction value; the acquired remote sensing image data is preprocessed to extract the terrain features, and the slope tool in the ArcGIS geographic information system is used. The slope tool processes the DEM elevation model data to generate a raster layer representing the slope. The generated raster image will display the slope value of each pixel, and the elevation change between each pixel and the surrounding 3×3 pixel neighborhood is calculated. By comparing the elevation values ​​of the central pixel with the adjacent pixels, The corresponding elevation change rate was determined to obtain the slope value, and the aspect tool in the ArcGIS geographic information system was used to generate an aspect raster layer. The aspect raster layer was reclassified and a numerical value was set for each category; 0° to 45°: category 1, with a value of 1; 45° to 90°: category 2, with a value of 2, 90° to 135°: category 3, with a value of 3, 135° to 180°: category 4, with a value of 4, 180° to 225°: category 5, with a value of 5, 225° to 2 70°: Category 6, value 6, 270° to 315°: Category 7, value 7, 315° to 360°: Category 8, value 8. Generate a reclassified aspect raster layer, extract ground height information from the DEM elevation model data, obtain the elevation value of each pixel and input it into the attribute table, use the raster calculator to extract the height value in the attribute table, standardize the values ​​and synthesize them to get the terrain complexity, synthesize the terrain complexity and the rainfall forecast value to get the terrain risk index; in, and are the weight coefficients of terrain risk index and rainfall prediction value, is 0.6, is 0.4.

[0025] In one embodiment of the present invention, the biometric recognition model is constructed using biological sample data obtained from historical landscapes, comprising the following steps: The biological sample data obtained from the historical landscape include the number and types of all species in the historical landscape. The plant, animal and microbial samples in the biological sample data are labeled and classified respectively. The tidyverse neat data science package in the R language is used to clean the data, including processing missing values, outliers and normalizing the data. The corresponding features of different types of biological samples are extracted through image processing software and text analysis methods. For plant samples, leaf shape, color and texture features are extracted. For animal samples, animal body shape, color and habitat features are extracted. For microbial samples, cell shape, size and color features are extracted. The statistical functions in the Python programming language are used to calculate the number of each type of biological samples. According to the features extracted from the data, the corresponding biometric recognition model is constructed, and the labeled data is used to train the model. The features extracted from the biological samples are used as input, and the types and corresponding quantities of biological samples are used as output. The trained model is deployed in a new environment to identify new biological samples.

[0026] Specifically, biological sample data are obtained from historical landscapes, and the collected samples are preliminarily processed, including cleaning, labeling, and classification. Missing values ​​and outliers are processed for the data after preliminary processing, and the data are normalized. Then, features are extracted from different types of biological samples, including plants, animals, and microorganisms, through image processing software. The number of each type of biological sample is calculated through statistical functions in the Python programming language, and the corresponding sample recognition model is constructed according to the features and quantity. The model is trained, and the trained model is deployed in a new environment to identify new biological samples.

[0027] In one embodiment of the present invention, the method of obtaining a comprehensive biological imbalance risk index according to the plant imbalance index, the animal imbalance index and the microbial imbalance index comprises the following steps: The species and quantity data of various biological samples obtained through the output of the biometric model are used to obtain the species uniformity decline index and species richness decline index of plants, animals and microorganisms respectively. The species imbalance index corresponding to different biological samples is obtained according to the average values ​​of the species uniformity decline index and species richness decline index of different biological samples respectively. The comprehensive biological imbalance risk index is obtained according to the plant imbalance index, animal imbalance index and microbial imbalance index. Species imbalance index calculation formula: in, is the species imbalance index, When is the plant imbalance index; When is the animal imbalance index; When is the microbial imbalance index; Any of plants, animals and microorganisms The number of individuals of a species, is the total number of individuals of any species of plants, animals and microorganisms, The total number of any one type of plant, animal and microorganism; Calculation formula for comprehensive biological imbalance risk index: in, is the comprehensive biological imbalance risk index, is the plant imbalance index, is the animal imbalance index, is the microbial imbalance index, , and are the corresponding weight coefficients.

[0028] Specifically, the species uniformity decline index and species richness decline index are obtained based on the characteristic data and quantity data of biological samples, and the two are averaged to construct the species imbalance index. Corresponding weights are assigned to the imbalance indexes of plants, animals and microorganisms, and the comprehensive biological imbalance risk index is obtained by adding them together. The increase in biological imbalance risk is often accompanied by a decrease in species richness, and it will interfere with the material cycle in the ecosystem, affect the stability of the entire ecosystem, and change the pattern of the historical landscape; in, , and are the weight coefficients of plant imbalance index, animal imbalance index and microbial imbalance index, respectively. is 0.4, is 0.4, is 0.2.

[0029] In one embodiment of the present invention, the air pollution risk index, terrain risk index and comprehensive biological imbalance risk index are respectively obtained by extracting characteristic data from the collected air environment monitoring data, topographic data and biological sample data to construct a model to obtain a comprehensive historical landscape risk assessment prediction value, including the following steps: The air pollution risk index, terrain risk index and comprehensive biological imbalance risk index are obtained by extracting characteristic data from the collected air environment monitoring data, topographic data and biological sample data. These data are used as the input features of the model, and the comprehensive historical landscape risk assessment prediction value is used as the output data to construct a historical landscape risk prediction model. The model is evaluated using the historical air pollution risk index, terrain risk index and comprehensive biological imbalance risk index.

[0030] Specifically, the corresponding data extracted from the collected air environment monitoring data, topographic data and biological sample data are calculated to obtain the air pollution risk index, terrain risk index and comprehensive biological imbalance risk index respectively. The air pollution risk index integrates the impact of various pollutants, the terrain risk index integrates the complex impact of various terrains, and the comprehensive biological imbalance risk index integrates various biological data to evaluate biodiversity. The above three risk indices are used as input features of the model to construct a historical landscape risk prediction model. The historical air pollution risk index, terrain risk index and comprehensive biological imbalance risk index are used to train the model. The trained model is deployed in the historical landscape environment for real-time risk analysis.

[0031] In one embodiment of the present invention, the setting of different risk thresholds and the classification and early warning of the risk levels of historical landscapes according to the predicted values ​​of the comprehensive historical landscape risk assessment include the following steps: Set different risk thresholds based on the distribution of historical data; If 0 < the predicted value of comprehensive historical landscape risk assessment < 0.2, it indicates a low-level risk level and a level 3 warning signal is issued; If 0.2≤the predicted value of the comprehensive historical landscape risk assessment<0.5, it indicates a medium risk level and a second-level warning signal is issued; If 0.5 ≤ the predicted value of the comprehensive historical landscape risk assessment, it indicates a high level of risk and a first-level warning signal is issued.

[0032] Specifically, the range of risk thresholds is determined based on the distribution of historical data. Within the determined threshold range, the risk level is refined. According to the size of the predicted value of the comprehensive historical landscape risk assessment, the risks are divided into three risk levels: low, medium and high. A corresponding early warning signal is set for each risk level. A third-level early warning signal is issued for the low-level risk level, a second-level early warning signal is issued for the medium risk level, and a first-level early warning signal is issued for the high-level risk level. The set risk thresholds and early warning signals are applied to the actual historical landscape.

[0033] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. Historical landscape risk assessment and early warning system based on big data, characterized by: It includes the following modules: Data collection and processing module: It collects various data, including air environment monitoring data, topographic data and biological sample data, through the combination of remote sensing technology and GIS technology, and pre-processes the data; Data analysis module: The relative offset values ​​of the corresponding pollutants obtained by comparing the measured values ​​of pollutants with the average values ​​of pollutant concentrations are used to obtain the air pollution risk index. The rainfall prediction model of the historical landscape is constructed through the historical rainfall data. The terrain complexity is obtained by comparing the rainfall prediction value with the terrain complexity to obtain the terrain risk index. The biological sample data obtained from the historical landscape are used to construct a biological recognition model. The comprehensive biological imbalance risk index is obtained based on the plant imbalance index, animal imbalance index and microbial imbalance index. Characteristic model prediction module: The characteristic data extracted from the collected air environment monitoring data, topographic data and biological sample data are used to obtain the air pollution risk index, topographic risk index and comprehensive biological imbalance risk index to build a model to obtain the comprehensive historical landscape risk assessment prediction value; Risk assessment and early warning module: Set different risk thresholds, and classify and warn the risk levels of historical landscapes based on the predicted values ​​of comprehensive historical landscape risk assessment.

2. The historical landscape risk assessment and early warning system based on big data according to claim 1 is characterized in that: The method combines remote sensing technology and GIS technology to collect a variety of data, including air environment monitoring data, topographic data and biological sample data, and pre-processes the data, including the following steps: after obtaining satellite image data using satellite sensors, use GPS technology to measure the position of ground control points, use GPS receivers to measure the coordinates of each ground control point, obtain lidar data through lidar technology to collect topographic data and process them to obtain a DEM digital elevation model, use satellite sensors to monitor pollutant types, and deploy air quality detection instruments in historical landscapes to collect real-time data, collect plants, animals and biological samples in historical landscape environments, use high-resolution images to identify and classify different biological species, and pre-process the collected data, including geometric correction and denoising.

3. The historical landscape risk assessment and early warning system based on big data according to claim 1 is characterized in that: The relative deviation values ​​of the corresponding pollutants obtained by comparing the measured values ​​of the pollutants with the average values ​​of the pollutant concentrations, and obtaining the air pollution risk index through the relative deviation values ​​of multiple pollutants, include the following steps: By deploying multiple air quality monitoring stations in the historical landscape, each station is equipped with a particle detector and a gaseous pollutant detector to collect data on PM2.5, SO2, NO2 and CO pollutants respectively, regularly monitor and record the corresponding concentration data of pollutants in the air, and after pre-processing the collected data, use big data analysis to obtain the average concentration of each pollutant. The relative offset value of the corresponding pollutant obtained by the pollutant measurement value collected by the measuring instrument and the average pollutant concentration is used to obtain the air pollution risk index through the relative offset values ​​of multiple pollutants; Air pollution risk index calculation formula: in, is the air pollution risk index, is the measured value of PM2.5 concentration in the air, is the average value of PM2.5 concentration, is the measured value of SO2 concentration in the air, is the average value of SO2 concentration, is the measured value of NO2 concentration in the air, is the average value of NO2 concentration, is the measured value of CO concentration in the air, is the average value of CO concentration.

4. The historical landscape risk assessment and early warning system based on big data according to claim 1 is characterized in that: The method of constructing a rainfall prediction model for historical landscapes through historical rainfall data, obtaining the terrain complexity from the feature data extracted from the remote sensing image data, and obtaining a terrain risk index through the rainfall prediction value and the terrain complexity, comprises the following steps: The remote sensing image data of the historical landscape is obtained through remote sensing technology, and the historical rainfall data of the historical landscape is obtained from the meteorological department. The time series features are extracted from the historical rainfall data using statistical feature analysis methods, including annual rainfall, average monthly rainfall and maximum daily rainfall. The extracted feature data is used as input variables, and the historical landscape rainfall data is used as output variables to train the model. The new remote sensing image data and feature data are input into the trained model. The model outputs the rainfall prediction value of the historical landscape based on the input data. The acquired remote sensing image data is preprocessed, including noise reduction, atmospheric correction and image enhancement. The terrain features, including slope, aspect and height features, are extracted from the preprocessed data. The slope tool in ArcGIS geographic information system is used to calculate the elevation change between each pixel and the surrounding 3×3 pixel neighborhood based on the DEM elevation model data. The corresponding elevation change rate is determined by comparing the elevation values ​​of the central pixel and the adjacent pixels to obtain the slope value; the aspect tool in ArcGIS geographic information system is used to calculate the value of each pixel based on the DEM elevation model data to represent the direction of the point. The aspect is classified by the reclassification tool, and a numerical value is set for each category to represent it; the ground height information is extracted from the DEM elevation model data, and the elevation value of each pixel is obtained and input into the attribute table. The height value is extracted from the attribute table using the raster calculator. The slope value, aspect value and height value are standardized and then combined to obtain the terrain complexity; The terrain complexity calculation formula is: in, is the complexity of the terrain, is the slope value, is the slope value, is the height value; The predicted rainfall values ​​of historical landscapes and the terrain complexity were normalized and integrated to obtain the terrain risk index; Terrain risk index calculation formula: in, is the terrain risk index, is the rainfall forecast value, and are all weight coefficients.

5. The historical landscape risk assessment and early warning system based on big data according to claim 1 is characterized in that: The method of constructing a biometric identification model using biological sample data obtained from historical landscapes comprises the following steps: obtaining biological sample data from historical landscapes, including the number and types of all species in the historical landscapes, labeling and classifying plant, animal and microbial samples in the biological sample data respectively, using the tidyverse neat data science package in the R language to clean the data, including processing missing values, outliers and normalizing the data, extracting corresponding features for different types of biological samples using image processing software and text analysis methods; for plant samples, extracting leaf shape, color and texture features; for animal samples, extracting animal body shape, color and habitat features; for microbial samples, extracting cell shape, size and color features, and using statistical functions in the Python programming language to calculate the number of each type of biological samples; constructing a corresponding biometric identification model based on the features extracted from the data, training the model using the labeled data, taking the features extracted from the biological samples as input, and taking the types and corresponding quantities of the biological samples as output, deploying the trained model in a new environment, and identifying new biological samples.

6. The historical landscape risk assessment and early warning system based on big data according to claim 1 is characterized in that: The method of obtaining a comprehensive biological imbalance risk index according to the plant imbalance index, the animal imbalance index and the microbial imbalance index comprises the following steps: The species and quantity data of various biological samples obtained through the output of the biometric model are used to obtain the species uniformity decline index and species richness decline index of plants, animals and microorganisms respectively. The species imbalance index corresponding to different biological samples is obtained according to the average values ​​of the species uniformity decline index and species richness decline index of different biological samples respectively. The comprehensive biological imbalance risk index is obtained according to the plant imbalance index, animal imbalance index and microbial imbalance index. Species imbalance index calculation formula: in, is the species imbalance index, When is the plant imbalance index; When is the animal imbalance index; When is the microbial imbalance index; Any of plants, animals and microorganisms The number of individuals of a species, is the total number of individuals of any species of plants, animals and microorganisms, The total number of any one type of plant, animal and microorganism; Calculation formula for comprehensive biological imbalance risk index: in, is the comprehensive biological imbalance risk index, is the plant imbalance index, is the animal imbalance index, is the microbial imbalance index, , and are the corresponding weight coefficients.

7. The historical landscape risk assessment and early warning system based on big data according to claim 1 is characterized in that: The air pollution risk index, terrain risk index and comprehensive biological imbalance risk index are respectively obtained by extracting characteristic data from the collected air environment monitoring data, topographic data and biological sample data to construct a model to obtain a comprehensive historical landscape risk assessment prediction value, including the following steps: The air pollution risk index, terrain risk index and comprehensive biological imbalance risk index are obtained by extracting characteristic data from the collected air environment monitoring data, topographic data and biological sample data. These data are used as the input features of the model, and the comprehensive historical landscape risk assessment prediction value is used as the output data to construct a historical landscape risk prediction model. The model is trained using the historical air pollution risk index, terrain risk index and comprehensive biological imbalance risk index. The trained model is deployed to the new historical landscape for real-time risk analysis.

8. The historical landscape risk assessment and early warning system based on big data according to claim 1 is characterized in that: The method of setting different risk thresholds and classifying and warning the risk levels of historical landscapes according to the predicted values ​​of the comprehensive historical landscape risk assessment includes the following steps: setting different risk thresholds according to the distribution of historical data; If 0 < the predicted value of comprehensive historical landscape risk assessment < 0.2, it indicates a low-level risk level and a level 3 warning signal is issued; If 0.2≤the predicted value of the comprehensive historical landscape risk assessment<0.5, it indicates a medium risk level and a second-level warning signal is issued; If 0.5 ≤ the predicted value of the comprehensive historical landscape risk assessment, it indicates a high level of risk and a first-level warning signal is issued.

Images