Method for quickly identifying natural disasters based on remote sensing satellite data

Abstract

The invention discloses a method for quickly identifying natural disasters based on remote sensing satellite data, and relates to the technical field of natural disaster monitoring. The problem that in the prior art, for complex and changeable natural disaster scenes, only a traditional data processing and analysis method is relied on, and the disaster type and degree are difficult to accurately and rapidly recognize is solved. According to the invention, rapid monitoring of a large area is realized by using multi-source remote sensing satellite data, a monitoring blind area is effectively eliminated, disasters in a remote area are found in time, the type and degree of the disasters can be accurately and rapidly identified, and the accuracy and reliability of disaster identification are greatly improved; according to the method, multi-source data are fully fused, and disaster information behind the data is deeply mined in combination with time sequence analysis and a machine learning algorithm, so that the data processing and analysis efficiency is remarkably improved, and the timeliness requirement of disaster emergency response is met; and a dynamically updated disaster feature library and an identification model are established, and the disaster identification and evaluation capability is continuously optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of natural disaster monitoring, and particularly to a method for quickly identifying natural disasters based on remote sensing satellite data. Background Art

[0002] Traditional methods for identifying natural disasters mainly rely on ground monitoring stations and manual inspections. The distribution of ground monitoring stations is limited, making it difficult to cover large areas and prone to monitoring blind spots, resulting in the inability to detect disasters in remote areas in a timely manner. The patent application with publication number CN118840829A discloses a method and system for evaluating the risk of geological disasters based on machine learning, including: determining the target area of geological disasters to be evaluated, obtaining multi-source data of the target area, processing the original data based on ArcGIS, dividing grid units, and extracting data on geological disaster-related factors for each grid unit; selecting evaluation indicators for the susceptibility of geological disasters according to the occurrence mechanism and regional characteristics of geological disasters; and determining evaluation indicators for the vulnerability of geological disasters; normalizing and standardizing the selected evaluation indicators to construct an evaluation index system for the risk of geological disasters, which can more accurately predict the risk of geological disasters and thus reduce the false alarm and missed alarm rates.

[0003] Although the above patent solves the problems of low efficiency of manual inspections, high consumption of manpower and material resources, and significant safety risks, there are still the following problems:

[0004] In the prior art, for complex and changeable natural disaster scenarios, it is difficult to accurately and quickly identify the type and degree of disasters only relying on traditional data processing and analysis methods, and the data acquisition and processing efficiency are low, which cannot meet the timeliness requirements of disaster emergency response, and the accuracy and reliability of disaster identification are limited. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for quickly identifying natural disasters based on remote sensing satellite data. Through multi-source data fusion, large-area and rapid monitoring can be achieved, monitoring blind spots can be eliminated, disasters in remote areas can be discovered in a timely manner, the timeliness requirements of emergency response can be met, the type and degree of disasters can be accurately identified, complex and changeable natural disaster scenarios can be adapted, and the accuracy and reliability of disaster identification can be significantly improved to solve the problems raised in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] A method for quickly identifying natural disasters based on remote sensing satellite data, including:

[0008] S1: Obtain multi-source remote sensing satellite data, and at the same time obtain geographical information data G data of the target area, and preprocess the obtained remote sensing satellite data;

[0009] S2: Perform change detection on the preprocessed multi-source remote sensing satellite data based on timestamps to identify the geographical information data G of the target area data of surface changes, determine the boundaries and scope of the surface changes, and mark them as disaster-affected areas;

[0010] S3: Extract the feature information of the disaster-affected areas from the preprocessed remote sensing satellite data, where the feature information includes spectral feature S spectral , texture feature S texture , shape feature S shape , and construct a disaster feature library D library ;

[0011] S4: Use the feature information of different types of natural disasters on remote sensing images as sample data S samples , input it into a pre-trained disaster recognition model for recognition, and obtain the recognition model M for each type of natural disaster model ;

[0012] S5: Based on the recognition model M for each type of natural disaster model judge whether natural disasters have occurred in the target area and the type of disaster D type , according to the judgment result, combined with the geographical information data G data and historical disaster data H data verify the recognition result;

[0013] S6: Calculate the recognition error rate of the disaster recognition model for the feature information in the remote sensing satellite data based on the verification result, and optimize and update the disaster feature library D library and the recognition model M for each type of natural disaster model ;

[0014] S7: Obtain the evaluation result after passing the verification, and conduct a disaster level assessment on the impact scope I range and severity I severity of natural disasters based on the disaster level division standard, and generate a natural disaster analysis report for the target area.

[0015] Furthermore, obtaining the multi-source remote sensing satellite data in S1 further includes: determining the geographical location range and time range of the target area based on the obtained geographical information data G of the target area data , extracting the meteorological data of the real-time target area; determining the timeliness requirements of the remote sensing satellite data based on the correlation between the meteorological data and natural disasters, selecting the corresponding remote sensing satellite data source in combination with the geographical location range of the target area, and obtaining the remote sensing satellite data from the selected remote sensing satellite data source.

[0016] Further, determining the boundary and scope of the surface change in S2 specifically includes:

[0017] Obtain the basic monitoring area boundary line B according to the actual coverage of the target area in multi-source remote sensing satellite data boundary ;

[0018] Based on the time stamp, perform change detection, extract the surface change information of the target area, and integrate the surface change information to generate a change data set D change ;

[0019] Identify the location P of the changed area from the change data set D change , calculate the linear distance value D between every two adjacent changed areas position , and evaluate the distribution and continuity of the surface change; distance Based on the geographic information data G of the target area

[0020] Analyze the terrain characteristics of the changed area, and determine the influence degree I of the terrain on the surface change according to the terrain characteristics of the changed area data ; impact ;

[0021] Compare the influence degree I impact with the preset influence degree threshold T threshold . If the influence degree I impact is lower than the preset influence degree threshold T threshold , it is determined that there is no significant influence;

[0022] Otherwise, it is determined that there is a significant influence, and adjust the basic monitoring area boundary line B according to the terrain characteristics of the changed area boundary ;

[0023] Finally, confirm and mark the boundary of the changed area. At the same time, calculate the geometric parameters of the changed area, and combine the analysis of the influence degree of the terrain characteristics and the change data set D change , and determine the scope of the changed area;

[0024] Mark the area corresponding to the determined boundary and scope of the surface change as the disaster area, and record the relevant attribute information of the disaster area at the same time.

[0025] Further, each surface change information in the change data set D change corresponds one-to-one with each change information time stamp T stamp to determine the specific time when each change information occurs.

[0026] Further, determining the influence degree I of the terrain on the surface change impact , specifically includes:

[0027] Build a terrain impact assessment model, take the terrain features of the changed area as input sample data, determine the standard change range of surface change information based on the terrain feature type, and extract the terrain features corresponding to each surface change information from the change dataset D change ;

[0028] Based on the time stamp T of each change information stamp Analyze the change characteristics of each surface change information, obtain the actual characteristic values of the change characteristics, analyze the change trend of each surface change information, and determine the weight of the terrain feature corresponding to each surface change information;

[0029] Input the weights of each terrain feature and the terrain feature data into the terrain impact assessment model for calculation to obtain the impact degree I of the terrain on the surface change impact .

[0030] Furthermore, in step S4, the feature information of different types of natural disasters on the remote sensing image is used as the sample data S samples , specifically including:

[0031] Preprocess the sample data S samples to obtain the preprocessed sample data S samples ;

[0032] Extract feature information from the preprocessed sample data S samples , and obtain the initial feature set of the sample data S samples according to the extraction result;

[0033] Retrieve the key features related to natural disaster identification from the initial feature set and integrate them into a key feature subset;

[0034] Obtain the shooting parameter information of the remote sensing image and the geographical information data G data of the target area, analyze the influence degree of different factors on the natural disaster feature expression according to the geographical information data G data and the remote sensing image shooting parameter information, and determine the feature weights of each key feature subset;

[0035] Based on the feature weights of each key feature subset and the preprocessed sample data S samples , calculate the basic feature values of the natural disaster features in each sample data S samples ;

[0036] If the sample data S samples contains multi-temporal remote sensing data, obtain the time series feature information corresponding to the sample data S samples , and analyze the change trend of natural disaster features at different time points.

[0037] Furthermore, obtain the sample data Ssamples The corresponding time series feature information further includes:

[0038] According to the time series feature information, extract the time series data of each sample from the sample data S samples to construct a time series data set;

[0039] Based on the time series data of each sample, combined with the basic feature values of the natural disaster features in each sample data S samples to determine the disaster feature description information corresponding to the sample;

[0040] Take the time series data and basic feature values of each sample as the model input samples, and at the same time take the disaster feature description information of each sample as the model output samples, and input them into a pre-set network model structure for training to obtain an identification model M for each type of natural disaster model .

[0041] Furthermore, the disaster level assessment results in S7 include level one, level two, level three, and level four; the natural disaster analysis report of the target area includes data on the basic situation, impact range, severity, danger level, and possible losses of the disaster.

[0042] Compared with the prior art, the beneficial effects of the present invention are:

[0043] Utilize multi-source remote sensing satellite data to achieve rapid monitoring of large areas, effectively eliminate monitoring blind spots, and promptly detect disasters in remote areas; through steps such as change detection, feature extraction, model training and verification, it is possible to accurately and quickly identify the type and degree of disasters, greatly improving the accuracy and reliability of disaster identification; fully integrate multi-source data, and combine time series analysis and machine learning algorithms to deeply mine the disaster information behind the data, significantly improving the data processing and analysis efficiency, and meeting the timeliness requirements of disaster emergency response; and establish a dynamically updated disaster feature library and identification model, which can adapt to changing natural disaster scenarios and continuously optimize the disaster identification and assessment capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is a flowchart of the method for quickly identifying natural disasters based on remote sensing satellite data of the present invention;

[0045] Figure 2 is a flowchart of step two of the present invention for change detection to determine the disaster-affected area. DETAILED DESCRIPTION OF THE INVENTION

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0047] To solve the technical problems in the prior art, such as low efficiency of data processing and analysis, insufficient multi-source data fusion, and difficulty in quickly and accurately identifying the type and degree of disasters, please refer to Figure 1-2 This embodiment provides the following technical solutions:

[0048] A method for quickly identifying natural disasters based on remote sensing satellite data, including:

[0049] S1: Obtain multi-source remote sensing satellite data, including optical satellite data O data , radar satellite data R data etc., and at the same time obtain the geographical information data G data of the target area, such as topographic and geomorphic data T geo , vegetation coverage data V cover etc., and preprocess the obtained remote sensing satellite data, including radiometric correction and geometric correction, to eliminate the influence of sensor errors and factors such as the earth's curvature and atmosphere; perform denoising and cropping processing, extract the affected area, and perform image fusion to improve the spatial resolution and spectral resolution and enhance the data quality;

[0050] In this embodiment, based on the obtained geographical information data G data of the target area, determine the geographical location range and time range of the target area, and extract the meteorological data of the real-time target area; determine the timeliness requirements of the remote sensing satellite data based on the correlation between the meteorological data and natural disasters. For example, for quickly occurring disasters (such as earthquakes and floods), the latest satellite data needs to be obtained, select the corresponding remote sensing satellite data source according to the geographical location range of the target area, and obtain the remote sensing satellite data from the selected remote sensing satellite data source. For example, for cloudy and rainy areas, the radar remote sensing image data source is preferably selected;

[0051] S2: Perform change detection on the preprocessed multi-source remote sensing satellite data based on the time stamp, identify the surface changes of the geographical information data G data of the target area, determine the boundary and range of the surface changes, and mark them as the affected area;

[0052] S3: Extract the feature information of the affected area from the preprocessed remote sensing satellite data, and the feature information includes spectral feature S spectral , texture feature S texture , shape feature Sshape and construct a disaster feature library D library ;

[0053] S4: Use the feature information of different types of natural disasters on remote sensing images as sample data S samples , input it into a pre-trained disaster recognition model for recognition, and obtain the recognition model M for each type of natural disaster model ;

[0054] S5: Based on the recognition model M for each type of natural disaster model judge whether a natural disaster has occurred in the target area and the type of disaster D type , according to the judgment result, combine the geographic information data G data and historical disaster data H data to verify the recognition result;

[0055] S6: Calculate the recognition error rate of the disaster recognition model for the feature information in the remote sensing satellite data based on the verification result, and optimize and update the disaster feature library D library and the recognition model M for each type of natural disaster model ;

[0056] S7: Obtain the evaluation result after passing the verification, and conduct a disaster level assessment on the impact range I range and severity I severity of natural disasters based on the disaster level division standard, generate a natural disaster analysis report for the target area, and the disaster level assessment result includes level one, level two, level three, and level four; the natural disaster analysis report for the target area includes data on the basic situation, impact range, severity, danger level, and possible losses of the disaster, providing a strong basis for disaster rescue, subsequent reconstruction, and prevention of similar disasters.

[0057] In this embodiment, for example, in the flood scenario, with the help of the location information of important facilities (such as city centers, important water conservancy hubs, etc.) in the geographic information data G data , according to the center position of the flood, calculate the straight-line distance between the flood center position and the nearest important facility, define this distance value as the first data, determine the geometric center position of the flood coverage area, calculate the straight-line distance between the center position of the flood coverage area and the nearest important facility, and this distance value is the second data. Construct a special risk scoring algorithm model, use the first data and the second data in the flood scenario as the key input parameters of the model, calculate the risk evaluation score of the flood, and pre-set three key thresholds for flood disasters as the division standard of the risk level. By comparing the risk evaluation score of the flood disaster with these preset thresholds, the risk level of the flood disaster is determined.

[0058] In this embodiment, by obtaining multi-source remote sensing satellite data and geographical information data of the target area and performing preprocessing, the data quality is ensured. At the same time, according to the real-time meteorological data and the disaster correlation relationship, the timeliness requirements of the remote sensing satellite data are determined, significantly improving the data processing and analysis efficiency, realizing the effective integration of multi-source data. Based on timestamp change detection, the disaster-affected areas are identified, characteristic information is extracted to construct a disaster feature library, the recognition model is trained and verified and optimized, the disaster type and degree can be quickly and accurately identified, and based on the disaster level division standard, the disaster impact scope and severity are evaluated, and an analysis report containing multi-dimensional data is generated, providing strong support for the full-process response to disasters, helping to reduce disaster losses and improve the disaster management level.

[0059] In this embodiment, in S2, determining the boundary and scope of the surface change specifically includes:

[0060] According to the actual coverage range of the target area in the multi-source remote sensing satellite data, obtain the basic monitoring area boundary line B boundary ;

[0061] Based on the timestamp, perform change detection, extract the surface change information of the target area, and integrate the surface change information to generate a change data set D change , including vegetation cover change V change , water area change W change , building damage B change etc.;

[0062] Identify the change area position P from the change data set D change , calculate the straight-line distance value D between every two adjacent change areas position , and evaluate the distribution and continuity of the surface change; distance Based on the geographical information data G of the target area

[0063] Analyze the terrain features of the change area, including slope S data , aspect A slope , terrain undulation degree R aspect etc. parameters, and determine the influence degree I of the terrain on the surface change according to the terrain features of the change area relief ; impact ;

[0064] Compare the influence degree I impact with the preset influence degree threshold T threshold . If the influence degree I impact is lower than the preset influence degree threshold T threshold , it is determined that there is no significant influence;

[0065] Otherwise, it is determined that there is a significant influence, and the basic monitoring area boundary line B is adjusted according to the terrain features of the change areaboundary , including buffer analysis based on terrain features, expanding a certain distance outward from the areas significantly affected by the terrain as the new boundary, or modifying the boundary according to the terrain change trend to adjust the boundary line B of the basic monitoring area boundary After that, it is necessary to re-extract the surface change information, identify the location of the change area, calculate the distance between change areas, etc., and re-analyze the change areas within the new boundary to ensure the accuracy of the boundary adjustment;

[0066] Finally, confirm and mark the boundary of the change area. At the same time, calculate geometric parameters such as the area and perimeter of the change area, and combine the impact degree analysis of terrain features and the change data set D change to determine the scope of the change area;

[0067] Mark the area corresponding to the determined boundary and scope of the surface change as the disaster-affected area. At the same time, record the relevant attribute information of the disaster-affected area, such as the change type (such as land cover change, terrain change, etc.), change time, change degree, etc., to provide detailed basis for disaster assessment and rescue decision-making.

[0068] In this embodiment, the change data set D change Each surface change information in it corresponds one-to-one with each change information timestamp T stamp to determine the specific time when each change information occurs, so as to analyze the time series characteristics of the change subsequently.

[0069] In this embodiment, by obtaining the boundary line of the basic monitoring area and performing change detection based on the timestamp, the surface change information can be accurately identified and recorded, thereby improving the accuracy of disaster identification. By calculating the linear distance value between adjacent change areas, the distribution and continuity of the surface change are evaluated, which helps to understand the diffusion pattern and influence range of the disaster. Combining the terrain features of the change area with geographical information data to deeply understand the impact of the terrain on the surface change, quantifying the impact of the terrain on the formation of the disaster, providing a quantitative index for disaster level assessment, and the dynamic boundary adjustment can more accurately define the disaster-affected area, providing detailed data support for disaster assessment and rescue decision-making.

[0070] In this embodiment, determining the influence degree I of the terrain on the surface change impact specifically includes:

[0071] Establish a terrain influence assessment model, use the terrain features of the change area as input sample data, determine the standard change range of the surface change information based on the terrain feature type, and extract the terrain features corresponding to each surface change information from the change data set D change ;

[0072] Based on each change information timestamp T stampAnalyze the change characteristics of each surface change information, obtain the actual characteristic values of the change characteristics, analyze the change trend of each surface change information, and determine the weight of the terrain characteristics corresponding to each surface change information;

[0073] Based on the weights of each terrain characteristic and the terrain characteristic data, input them into the terrain impact assessment model for calculation to obtain the impact degree I of the terrain on the surface change impact 。

[0074] In this embodiment, by establishing a terrain impact assessment model, combining multi-source remote sensing data and terrain characteristics, accurately calculate the impact degree of the terrain on the surface change, improve the scientificity and response efficiency of disaster identification, analyze the change trend and dynamically allocate characteristic weights, provide comprehensive data support, and effectively reduce disaster losses.

[0075] In this embodiment, in S4, the characteristic information of different types of natural disasters on the remote sensing image is used as the sample data S samples ,specifically including:

[0076] Preprocess the sample data S samples to obtain the processed sample data S samples ;

[0077] Extract characteristic information from the processed sample data S samples including spectral characteristics, texture characteristics, shape characteristics, etc., and obtain the initial characteristic set of the sample data S samples according to the extraction results;

[0078] Retrieve the key characteristics related to natural disaster identification from the initial characteristic set and integrate them into a key characteristic subset;

[0079] Obtain the shooting parameter information of the remote sensing image, such as sensor type, imaging time, imaging angle, etc., and the geographical information data G of the target area data ,and analyze the influence degree of different factors on the expression of natural disaster characteristics according to the geographical information data G data and the remote sensing image shooting parameter information, and determine the characteristic weights of each key characteristic subset. For example, in mountainous areas, the terrain and landform have a greater impact on the characteristic performance of disasters such as landslides and debris flows, and the corresponding terrain characteristic weights can be increased accordingly; in different seasons, the vegetation growth conditions are different, and the weights of vegetation-related characteristics in the identification of some disasters should also be adjusted accordingly;

[0080] Based on the characteristic weights of each key characteristic subset and the processed sample data S samples ,calculate the basic characteristic values of the natural disaster characteristics in each sample data S samples ;

[0081] If the sample data Ssamples Including multi-temporal remote sensing data, obtain sample data S samples The corresponding time series feature information, and analyze the change trend of natural disaster characteristics at different time points.

[0082] In this embodiment, obtain sample data S samples The corresponding time series feature information further includes:

[0083] According to the time series feature information, extract the time series data of each sample from the sample data S samples to construct a time series data set;

[0084] Based on the time series data of each sample, combined with the basic feature values of the natural disaster characteristics in each sample data S samples determine the disaster feature description information corresponding to the sample. For example, analyze the change amplitude, change frequency, etc. of the disaster characteristics in the time series data, and combine them with the basic feature values to comprehensively describe the natural disaster characteristics represented by each sample;

[0085] Take the time series data and basic feature values of each sample as the model input samples, and at the same time take the disaster feature description information of each sample as the model output samples, and input them into a pre-set network model structure for training to obtain an identification model M for each type of natural disaster model .

[0086] In this embodiment, through refined feature extraction and weight assignment, combined with time series analysis, an identification model for different types of natural disasters is constructed, which improves the accuracy of feature recognition and the timeliness of disaster warning, can more comprehensively describe the development process of natural disasters, enhances the ability of dynamic disaster monitoring, can effectively improve the identification efficiency and accuracy of natural disasters, provides strong technical support for pre-disaster warning, post-disaster rescue and risk assessment, and significantly improves the overall effectiveness of disaster prevention and mitigation.

[0087] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A method for quickly identifying natural disasters based on remote sensing satellite data, characterized in that: include: S1: Acquire multi-source remote sensing satellite data and simultaneously acquire geographic information data G of the target area data , and pre-process the acquired remote sensing satellite data; S2: Perform change detection on the preprocessed multi-source remote sensing satellite data based on timestamps to identify the geographic information data G of the target area data Surface changes, determine the boundaries and scope of surface changes, and mark them as disaster-affected areas; S3: Extract characteristic information of the disaster-affected area from the pre-processed remote sensing satellite data, the characteristic information includes spectral characteristics S spectral , texture feature S texture , shape feature S shape , and build a disaster feature database D library ; S4: The characteristic information of different types of natural disasters on remote sensing images is used as sample data S samples , input into the pre-trained disaster identification model for identification, and obtain the identification model M for each type of natural disaster model ; S5: Identification model M based on each type of natural disaster model Determine whether a natural disaster has occurred in the target area and the type of disaster D type , based on the judgment results, combined with geographic information data G data and historical disaster data H data Verify the recognition results; S6: Based on the verification results, calculate the recognition error rate of the disaster identification model for the feature information in the remote sensing satellite data, and use the recognition error rate to identify the disaster feature database D. library and the identification model M for each type of natural disaster model Carry out optimization and update; S7: Obtain the verified assessment results and the impact range of natural disasters based on the disaster level classification standards I range and severity level I severity Conduct disaster level assessment and generate natural disaster analysis reports for target areas.

2. The method for rapidly identifying natural disasters based on remote sensing satellite data as claimed in claim 1, characterized in that: The step of obtaining multi-source remote sensing satellite data in S1 further includes: obtaining geographic information data G of the target area based on the obtained data Determine the geographical location range and time range of the target area, and extract real-time meteorological data of the target area; determine the timeliness requirements of remote sensing satellite data based on the correlation between meteorological data and natural disasters, select the corresponding remote sensing satellite data source in combination with the geographical location range of the target area, and obtain remote sensing satellite data from the selected remote sensing satellite data source.

3. The method for rapidly identifying natural disasters based on remote sensing satellite data as claimed in claim 2, characterized in that: The boundary and scope of the surface change determined in S2 specifically include: According to the actual coverage of the target area in the multi-source remote sensing satellite data, the basic monitoring area boundary line B is obtained. boundary ; Perform change detection based on timestamps, extract surface change information of the target area, and integrate the surface change information to generate a change dataset D change ; From the variation data set D change The position of the changed area P is identified in position , calculate the straight-line distance D between each two adjacent change areas distance , assess the distribution and continuity of land surface changes; Based on the geographic information data G of the target area data Analyze the terrain characteristics of the changing area and determine the degree of influence of the terrain on the surface change based on the terrain characteristics of the changing area. impact ; Will affect the degree I impact and the preset impact threshold T threshold Compare, if the impact level is I impact Lower than the preset impact threshold T threshold , then it is judged that there is no significant effect; Otherwise, it is judged that there is a significant impact, and the basic monitoring area boundary line B is adjusted according to the terrain characteristics of the changed area. boundary ; Finally, the boundaries of the changed area are confirmed and marked. At the same time, the geometric parameters of the changed area are calculated, combined with the impact analysis of terrain features and the changed data set D change , determine the scope of the change area; The area corresponding to the boundary and range of the determined surface changes is marked as the disaster-affected area, and at the same time, the relevant attribute information of the disaster-affected area is recorded.

4. The method for rapidly identifying natural disasters based on remote sensing satellite data as claimed in claim 3, characterized in that: The change data set D change Each surface change information and each change information timestamp T stamp One-to-one correspondence determines the specific time when each change information occurs.

5. The method for rapidly identifying natural disasters based on remote sensing satellite data as claimed in claim 4, characterized in that: Determine the extent to which topography affects surface changes I impact , including: A terrain impact assessment model is established, and the terrain characteristics of the change area are used as input sample data. The standard change range of surface change information is determined based on the type of terrain characteristics, and the standard change range of the surface change information is determined from the change data set D. change Extract the terrain features corresponding to each surface change information; Based on each change information timestamp T stamp Analyze the change characteristics of each surface change information, obtain the actual characteristic value of the change characteristics, analyze the change trend of each surface change information, and determine the weight of the terrain feature corresponding to each surface change information; The weights of the terrain features and the terrain feature data are input into the terrain impact assessment model for calculation to obtain the impact of the terrain on the surface change. impact .

6. The method for rapidly identifying natural disasters based on remote sensing satellite data as claimed in claim 5, characterized in that: In S4, the characteristic information of different types of natural disasters on remote sensing images is used as sample data S samples , including: For the sample data S samples Perform preprocessing to obtain processed sample data S samples ; From the processed sample data S samples Extract feature information and obtain sample data S according to the extraction results samples The initial feature set of Retrieving key features related to natural disaster identification from the initial feature set and integrating them into a key feature subset; Obtain the shooting parameter information of the remote sensing image and the geographic information data G of the target area data , according to geographic information data G data The influence of different factors on the expression of natural disaster characteristics is analyzed based on the remote sensing image shooting parameter information, and the feature weight of each key feature subset is determined; Based on the feature weights of each key feature subset combined with the processed sample data S samples , calculate each sample data S samples Basic characteristic values ​​of natural disaster characteristics in China; If the sample data S samples Contains multi-temporal remote sensing data, obtains sample data S samples Corresponding time series characteristic information is used to analyze the changing trends of natural disaster characteristics at different time points.

7. The method for rapidly identifying natural disasters based on remote sensing satellite data as claimed in claim 6, characterized in that: Get sample data S samples The corresponding time series feature information also includes: According to the time series characteristic information, from the sample data S samples Extract the time series data of each sample and construct a time series dataset; Based on the time series data of each sample, combined with each sample data S samples The basic characteristic value of the natural disaster characteristics in the sample is used to determine the disaster characteristic description information corresponding to the sample; The time series data and basic feature values ​​of each sample are used as the model input samples, and the disaster feature description information of each sample is used as the model output sample, which is input into the pre-set network model structure for training to obtain the recognition model M for each type of natural disaster. model .

8. The method for rapidly identifying natural disasters based on remote sensing satellite data as claimed in claim 7, characterized in that: The disaster level assessment results in S7 include level one, level two, level three and level four; the natural disaster analysis report of the target area includes data on the basic situation of the disaster, the scope of impact, severity, dangerousness level and possible losses.

Images