Seismic landslide risk zoning method based on anti-clustering ANFIS

Abstract

The invention discloses a seismic landslide risk zoning method based on a subclustering ANFIS, and relates to the technical field of seismic landslide risk zoning, and the method comprises the following steps: collecting seismic landslide monitoring data including seismic data, landform data, meteorological and hydrological data and image data, carrying out the feature extraction of the image data, and obtaining feature extraction data; based on the seismic landslide monitoring data, an initial clustering center point of a seismic landslide area is determined by adopting a subclustering algorithm, and then seismic landslide initial clustering information is acquired; according to the method, a data acquisition technology, a subclustering algorithm application technology, an ANFIS model construction technology and a data calibration technology in the method are closely combined with a modern information technology, and the data acquisition technology, the subclustering algorithm application technology, the ANFIS model construction technology, the data calibration technology and the modern information technology are combined together, so that the earthquake landslide danger level and the earthquake landslide monitoring data in each geographical grid unit are combined together; and the risk index is calibrated by using the risk calibration coefficient, so that the earthquake landslide risk assessment precision is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of seismic landslide hazard zoning, and particularly to a seismic landslide hazard zoning method based on subtractive clustering ANFIS. Background Art

[0002] As a highly destructive geological disaster, seismic landslides pose a serious threat to human life and property. In areas with frequent plate activities, earthquakes and landslides often occur concomitantly, causing huge losses to the local area. In the early stage, the assessment of seismic landslide hazards mostly relied on empirical qualitative analysis and simple mathematical models, making it difficult to accurately consider the complex interactions of multiple factors such as earthquakes, topography, geology, and meteorology. The accuracy and reliability of the assessment results were poor. With the rise of GIS technology and machine learning technology, it provided technical support for improving the accuracy of seismic landslide hazard assessment, fully exploited the potential information of multi-source data of seismic landslides, accurately identified seismic landslide risks, provided a more targeted and reliable basis for seismic landslide prevention and control and planning, and provided a solution for reducing the losses caused by seismic landslides. Although the existing technology has made great progress in the direction of seismic landslide hazard zoning, there are still some problems to be optimized. The traditional seismic landslide hazard zoning method relies on historical data samples and has high requirements for the integrity and accuracy of historical data. When facing complex geological conditions and seismic interactions, it is difficult to accurately describe the non-linear relationship. Moreover, there are also limitations in realizing seismic landslide hazard zoning through machine learning methods. Therefore, how to combine the subtractive clustering algorithm with the ANFIS model to solve the limitations of the traditional seismic landslide hazard zoning method and achieve the zoning of seismic landslide hazards. For this purpose, a seismic landslide hazard zoning method based on subtractive clustering ANFIS is proposed. Summary of the Invention

[0003] To achieve the above objectives, the present invention is realized through the following technical solutions: A seismic landslide hazard zoning method based on subtractive clustering ANFIS, comprising the following steps: Step 1: Collect seismic landslide monitoring data including seismic data, topographic and geomorphic data, meteorological and hydrological data, and image data, providing a data basis for the implementation of subsequent steps; Step 2: Extract features from the image data to obtain feature extraction data, laying a foundation for correcting the seismic landslide hazard index; Step 3: Based on the seismic landslide monitoring data, use the subtractive clustering algorithm to determine the initial clustering center points of the seismic landslide area, and then obtain the initial clustering information of the seismic landslide, providing data preparation for constructing the ANFIS model; Step 4: Based on the obtained initial clustering information of seismic landslides, construct an ANFIS model, and train the ANFIS model to obtain the final subtractive clustering ANFIS model; Step 5: Divide the seismic landslide monitoring area into individual geographical grids, assign corresponding seismic landslide monitoring data to each geographical grid unit, and combine the subtractive clustering ANFIS model to obtain the seismic landslide hazard index of each geographical grid unit; Step 6: Combine meteorological and hydrological data, feature extraction data, and neural network algorithms to construct a seismic landslide calibration model, and then obtain the hazard calibration coefficient; Step 7: Use the hazard calibration coefficient to calibrate the seismic landslide hazard index of each geographical grid unit, and divide the seismic landslide hazard levels within each geographical grid unit; Step 8: Combine the seismic landslide hazard levels within each geographical grid unit and their seismic landslide monitoring data to draw a seismic landslide hazard zoning map.

[0004] A further improvement of the technical solution of the present invention is that in the above Step 1, the process of collecting seismic landslide monitoring data includes: Deploy different types of collection devices, and combine geological monitoring stations, collection devices, and regional databases to collect seismic data, topographic and geomorphic data, meteorological and hydrological data, and image data of the seismic landslide monitoring area. Among them, the collection devices include total stations, altitude measuring instruments, temperature sensors, ultrasonic anemometers, optical rain gauges, float type water level gauges, electromagnetic flow meters, sediment concentration measuring instruments, electromagnetic current meters, and optical cameras; Specifically, collect seismic data through geological monitoring stations, and collect topographic and geomorphic data, meteorological and hydrological data, and image data through collection devices; The seismic data includes seismic intensity, seismic frequency, seismic duration, seismic peak acceleration, and epicentral distance; the topographic and geomorphic data includes vegetation coverage, lithology, soil composition, slope, altitude, plane curvature, and profile curvature; the meteorological and hydrological data includes air temperature, wind speed, precipitation, and the water level, flow rate, sediment concentration, and flow velocity of surrounding water bodies; the image data is a remote sensing image of the seismic landslide monitoring area; Perform data cleaning and data standardization processing on the collected seismic data, topographic and geomorphic data, and meteorological and hydrological data, perform image enhancement and image denoising processing on the collected image data, integrate the preprocessed seismic data, topographic and geomorphic data, and meteorological and hydrological data to generate a seismic landslide monitoring data set, and divide the seismic landslide monitoring data set into a training set and a test set, and the ratio of the training set to the test set is 7:3.

[0005] A further improvement of the technical solution of the present invention is that in the above Step 2, the process of extracting feature data includes: Import the remote sensing images of the earthquake landslide monitoring area into ENVI software. Using RNVI software, extract the gray values of the green light band and the near-infrared band from the remote sensing images of the earthquake landslide monitoring area respectively. Then, convert the gray values of the green light band and the near-infrared band into the reflectance values of the green light band and the near-infrared band respectively; Based on the reflectance values of the green light band and the near-infrared band, use the multi-spectral index method to calculate the normalized difference water index (NDWI). Its calculation formula is: ; Set the normalized difference water index threshold, and based on the normalized difference water index threshold, identify the water area and non-water area within the earthquake landslide monitoring area; Through GIS software, obtain the coordinates of each point within the earthquake landslide monitoring area. Starting from the boundary points of the earthquake landslide monitoring area and ending at the boundary points of the water body, combined with the Euclidean distance formula, calculate the distance from the starting point to the ending point, which is the distance between the earthquake landslide monitoring area and the surrounding water area. Then, use it as the feature extraction data and integrate the obtained feature extraction data into the earthquake landslide monitoring dataset. The calculation process is as follows: where, is the distance between the earthquake landslide monitoring area and the surrounding water area; ( , ) are the coordinates of the boundary points of the earthquake landslide monitoring area, that is, the starting point coordinates; ([[]] , ) are the coordinates of the boundary points of the water body, that is, the ending point coordinates.

[0006] A further improvement of the technical solution of the present invention lies in that: in the third step, the process of obtaining the initial clustering information of the earthquake landslide includes: The initial clustering information of the earthquake landslide includes seismic data, topographic and geomorphic data, and meteorological and hydrological data of the initial clustering center point of the earthquake landslide; Set the earthquake landslide monitoring data points, divide the earthquake landslide monitoring area into several small monitoring areas, and represent each small monitoring area through each earthquake landslide monitoring data point; Based on the seismic data, topographic and geomorphic data, and meteorological and hydrological data of the earthquake landslide monitoring area, obtain the seismic data, topographic and geomorphic data, and meteorological and hydrological data corresponding to the earthquake landslide monitoring data points; Integrate the seismic data, topographic and geomorphic data, and meteorological and hydrological data within each earthquake landslide monitoring data point to generate a feature vector of the earthquake landslide monitoring data point; Through the subtractive clustering algorithm, calculate the density of each earthquake landslide monitoring data point, and take the earthquake landslide monitoring data point with the maximum density as the first clustering center point of the earthquake landslide, and update the density of the earthquake landslide monitoring data points; Compare the density of the updated earthquake landslide monitoring data points with the density of the first clustering center point of the earthquake landslide. If the density of the updated earthquake landslide monitoring data points is less than the density of the first clustering center point of the earthquake landslide, then take the first clustering center point of the earthquake landslide as the initial clustering center point of the earthquake landslide; if the density of the updated earthquake landslide monitoring data points is greater than the density of the first clustering center point of the earthquake landslide, then take the updated earthquake landslide monitoring data points as the initial clustering center point of the earthquake landslide, and repeat updating the density of the earthquake landslide monitoring data points until the earthquake landslide monitoring data points with the maximum density are obtained, and take them as the initial clustering center point of the earthquake landslide; The calculation process of the density of the earthquake landslide monitoring data points and the density of the updated earthquake landslide monitoring data points includes: Among them, is the density of the earthquake landslide monitoring data points; is the density of the updated earthquake landslide monitoring data points; and are the feature vectors of the earthquake landslide monitoring data point i and the earthquake landslide monitoring data point j respectively; and are the influence range parameters; is the feature vector of the h-th clustering center point; is the density value of the h-th clustering center point; Through the feature vectors of the earthquake landslide monitoring data points, the earthquake data, topographic and geomorphic data, and meteorological and hydrological data of the initial clustering center point of the earthquake landslide are correspondingly obtained.

[0007] A further improvement of the technical solution of the present invention lies in that: in the fourth step, the construction process of the ANFIS model includes: Take the initial clustering information of the earthquake landslide as the input variable of the ANFIS model, define the fuzzy membership function and fuzzy set for each input variable, substitute the actual input variable data into the corresponding fuzzy membership function, and convert each input variable into a fuzzy membership value to realize the fuzzification of the input data; Specifically, for earthquake intensity, the corresponding fuzzy sets are defined as low intensity, medium intensity, and high intensity; for earthquake frequency, the corresponding fuzzy sets are defined as low frequency, medium frequency, and high frequency; for earthquake duration, the corresponding fuzzy sets are defined as very short, short, medium, long, and very long; for peak ground acceleration, the corresponding fuzzy sets are defined as very small, small, medium, large, and very large; for epicentral distance, the corresponding fuzzy sets are defined as very close, close, relatively close, far, and very far; for slope, the corresponding fuzzy sets are defined as very gentle, gentle, relatively steep, steep, and very steep; for elevation, the corresponding fuzzy sets are defined as very low, low, medium, high, and very high; for planar curvature, the corresponding fuzzy sets are defined as very concave, concave, near horizontal, convex, and very convex; for profile curvature, the corresponding fuzzy sets are defined as very concave, concave, near horizontal, convex, and very convex; for air temperature, the corresponding fuzzy sets are defined as very cold, cold, moderate, hot, and very hot; for wind speed, the corresponding fuzzy sets are defined as very small, small, medium, large, and very large; for rainfall, the corresponding fuzzy sets are defined as light rain, moderate rain, heavy rain, and rainstorm; for water level, the corresponding fuzzy sets are defined as very low, relatively low, medium, high, and very high; for water discharge, the corresponding fuzzy sets are defined as very small, small, medium, large, and very large; for sediment concentration, the corresponding fuzzy sets are defined as very little, little, medium, much, and very much; for water velocity, the corresponding fuzzy sets are defined as very slow, slow, medium speed, fast, and very fast; Based on the initial clustering center points of earthquake landslides obtained by the subtractive clustering algorithm, determine the number of fuzzy rules. Through the logical combination between each input variable and the corresponding fuzzy set, using the AND relationship and the OR relationship, construct a fuzzy rule base; According to the fuzzy rule base, perform inference operations on the fuzzy results of the input variables, obtain the output results of each fuzzy rule, and synthesize the outputs of all fuzzy rules to obtain the total fuzzy rule output result; Through the maximum membership degree method, defuzzify the total fuzzy rule output result, select the domain value corresponding to the maximum membership degree as the output value of the ANFIS model, and then construct the ANFIS model. The domain value corresponding to the maximum membership degree is the interval of the fuzzy membership function when the membership degree of the input variable in the fuzzy set reaches the maximum value. The output value of the ANFIS model is the earthquake landslide hazard index.

[0008] A further improvement of the technical solution of the present invention lies in that: in the fourth step, the process of training the ANFIS model and obtaining the subtractive clustering ANFIS model includes: Extract the earthquake data, topographic and geomorphic data, and meteorological and hydrological data from the earthquake landslide monitoring dataset. Using the training set data, combined with the backpropagation algorithm, take the earthquake data, topographic and geomorphic data, and meteorological and hydrological data as inputs and the earthquake landslide hazard index as the output, learn the non-linear relationship between the earthquake data, topographic and geomorphic data, meteorological and hydrological data and the earthquake landslide hazard index, and train the ANFIS model; Input the test set data into the ANFIS model, compare the output results of the ANFIS model with the actual seismic landslide hazard index, adjust the parameters of the ANFIS model, optimize the ANFIS model, and obtain the subtractive clustering ANFIS model.

[0009] A further improvement of the technical solution of the present invention lies in that: in step five, the process of obtaining the seismic landslide hazard index of each geographical grid unit includes: According to the small monitoring areas corresponding to the seismic landslide monitoring data points, divide the seismic landslide monitoring area into each geographical grid, and then obtain the seismic landslide monitoring data of each geographical grid unit; Combined with seismic data, topographic and geomorphic data, meteorological and hydrological data and the subtractive clustering ANFIS model, obtain the seismic landslide hazard index of each geographical grid unit.

[0010] A further improvement of the technical solution of the present invention lies in that: in step six, the process of obtaining the hazard calibration coefficient includes: Extract the feature extraction data in the seismic landslide monitoring data set, the water level, flow rate, sediment content and flow velocity of the water body around the seismic landslide monitoring area, and comprehensively analyze the vegetation coverage, lithology and soil composition of the monitoring area as input parameters, integrate the extracted input parameters, and construct an input data set; Combined with the input data set and the multiple linear regression algorithm, take the extracted input parameters as input data, take the hazard calibration coefficient as output data, learn the linear relationship between the input data and the output data, and train the seismic landslide calibration model; Input the input data set into the seismic landslide calibration model, adjust the intercept term and regression coefficient of the seismic landslide calibration model, optimize the performance of the seismic landslide calibration model, deploy the optimized seismic landslide calibration model into the system, and combine the current input parameters to output the corresponding hazard calibration coefficient; The expression of the seismic landslide calibration model is: Among them, is the hazard calibration coefficient; , , , are the regression coefficients of each input parameter respectively; , , are each input parameter respectively; and are the intercept term and error term of the seismic landslide calibration model respectively.

[0011] A further improvement of the technical solution of the present invention lies in that: in the seventh step, the process of calibrating the seismic landslide hazard index of each geographical grid unit and dividing the seismic landslide hazard levels within each geographical grid unit includes: Using the hazard calibration coefficient to calibrate the seismic landslide hazard index of each geographical grid unit to obtain the calibrated seismic landslide hazard index of each geographical grid unit; Set the first threshold of the seismic landslide hazard index and the second threshold of the seismic landslide hazard index. When the seismic landslide hazard index of each geographical grid unit is lower than the first threshold of the seismic landslide hazard index, the corresponding geographical grid unit is of a low seismic landslide hazard level; when the seismic landslide hazard index of each geographical grid unit is between the first threshold and the second threshold of the seismic landslide hazard index, the corresponding geographical grid unit is of a medium seismic landslide hazard level; when the seismic landslide hazard index of each geographical grid unit is higher than the first threshold of the seismic landslide hazard index, the corresponding geographical grid unit is of a high seismic landslide hazard level.

[0012] A further improvement of the technical solution of the present invention lies in that: in the eighth step, the process of drawing the seismic landslide hazard zoning map includes: Import the seismic landslide hazard levels within each geographical grid unit into GIS software. Using the GIS software, set green icons, yellow icons, and red icons for the low seismic landslide hazard level, medium seismic landslide hazard level, and high seismic landslide hazard level respectively, and then obtain the seismic landslide hazard level layer; Convert the seismic data, topographic and geomorphic data, and meteorological and hydrological data of each geographical grid unit into raster data, and perform visualization settings on the raster data to obtain the seismic data layer, topographic and geomorphic data layer, and meteorological and hydrological data layer respectively; Overlay the seismic landslide hazard level layer, seismic data layer, topographic and geomorphic data layer, and meteorological and hydrological data layer to draw the seismic landslide hazard zoning map.

[0013] The beneficial effects of the present invention are as follows: A seismic landslide hazard zoning method based on subtractive clustering ANFIS in the present invention, compared with traditional seismic landslide hazard assessment methods, the data acquisition technology, subtractive clustering algorithm application technology, ANFIS model construction technology, and data calibration technology in the method of the present invention are closely combined with modern information technology to accurately capture seismic data, topographic and geomorphic data, meteorological and hydrological data, and image data, and then obtain initial clustering information of seismic landslides, hazard indexes, and hazard calibration coefficients. Based on the initial clustering information of seismic landslides, an ANFIS model is constructed, and the hazard indexes are calibrated using the hazard calibration coefficients, thus achieving a more accurate seismic landslide hazard assessment, realizing effective monitoring in the process of seismic landslide hazard assessment, solving the problem that the traditional seismic landslide hazard zoning method depends on historical data samples and has high requirements for the integrity and accuracy of historical data, and is difficult to solve the non-linear relationship between complex geological conditions and seismic actions, ensuring that the method in the present invention can refine the dynamic monitoring standard of seismic landslide hazards within a more accurate range, making the monitored data a more accurate index under the same conditions. The research and application of this method significantly enhance the degree of intelligence in the process of seismic landslide hazard assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0015] Figure 1 It is a flowchart of a seismic landslide hazard zoning method based on subtractive clustering ANFIS of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0017] As Figure 1 shown, the present invention provides a seismic landslide hazard zoning method based on subtractive clustering ANFIS, which consists of the following steps: Step 1: Collect seismic landslide monitoring data including seismic data, topographic and geomorphic data, meteorological and hydrological data, and image data, providing a data basis for the implementation of subsequent steps; Step 2: Extract features from the image data to obtain feature extraction data, which lays the foundation for calibrating the seismic landslide hazard index; Step 3: Based on the seismic landslide monitoring data, use the subtractive clustering algorithm to determine the initial clustering center points in the seismic landslide area, and then obtain the initial clustering information of the seismic landslide, providing data preparation for constructing the ANFIS model; Step 4: Based on the obtained initial clustering information of the seismic landslide, construct an ANFIS model and train the ANFIS model to obtain the final subtractive clustering ANFIS model; Step 5: Divide the seismic landslide monitoring area into individual geographical grids, assign corresponding seismic landslide monitoring data to each geographical grid unit, and combine with the subtractive clustering ANFIS model to obtain the seismic landslide hazard index of each geographical grid unit; Step 6: Combine the meteorological and hydrological data, feature extraction data, and neural network algorithm to construct a seismic landslide calibration model, and then obtain the hazard calibration coefficient; Step 7: Use the hazard calibration coefficient to calibrate the seismic landslide hazard index of each geographical grid unit and divide the seismic landslide hazard levels within each geographical grid unit; Step 8: Combine the seismic landslide hazard levels and their seismic landslide monitoring data within each geographical grid unit to draw a seismic landslide hazard zoning map.

[0018] In Step 1, the process of collecting seismic landslide monitoring data includes: Deploy different types of collection devices, and combine geological monitoring stations, collection devices, and regional databases to collect seismic data, topographic and geomorphic data, meteorological and hydrological data, and image data in the seismic landslide monitoring area. Among them, the collection devices include total stations, altitude measuring instruments, temperature sensors, ultrasonic anemometers, optical rain gauges, float-type water level gauges, electromagnetic flowmeters, sediment concentration measuring instruments, electromagnetic current meters, and optical cameras; Specifically, collect seismic data through geological monitoring stations; collect topographic and geomorphic data, meteorological and hydrological data, and image data through collection devices; Specifically, seismic data includes seismic intensity, seismic frequency, seismic duration, seismic peak acceleration, and epicentral distance; topographic and geomorphic data includes vegetation coverage, lithology, soil composition, slope, altitude, planar curvature, and profile curvature; the meteorological and hydrological data includes air temperature, wind speed, precipitation, and the water level, flow rate, sediment concentration, and flow velocity of surrounding water bodies; the image data is a remote sensing image of the seismic landslide monitoring area; Clean and standardize the collected seismic data, topographic and geomorphic data, and meteorological and hydrological data, enhance and denoise the collected image data, integrate the preprocessed seismic data, topographic and geomorphic data, and meteorological and hydrological data to generate a seismic landslide monitoring dataset, and divide the seismic landslide monitoring dataset into a training set and a test set, with the ratio of the training set to the test set being 7:3.

[0019] In step two, the process of extracting feature data includes: Specifically, import the remote sensing image of the seismic landslide monitoring area into ENVI software, and use RNVI software to extract the gray values of the green light band and the near-infrared band from the remote sensing image of the seismic landslide monitoring area, and then convert the gray values of the green light band and the near-infrared band into the reflectance values of the green light band and the near-infrared band respectively; Based on the reflectance values of the green light band and the near-infrared band, use the multispectral index method to calculate the normalized difference water index, and its calculation formula is: ; Set the normalized difference water index threshold, and based on the normalized difference water index threshold, identify the water area and non-water area within the seismic landslide monitoring area; Through GIS software, obtain the coordinates of each point within the seismic landslide monitoring area, starting from the boundary point of the seismic landslide monitoring area and ending at the boundary point of the water body, and combine the Euclidean distance formula to calculate the distance from the starting point to the ending point, which is the distance between the seismic landslide monitoring area and the surrounding water area, and then use it as the feature extraction data, and integrate the obtained feature extraction data into the seismic landslide monitoring dataset. The calculation process is as follows: Among them, is the distance between the seismic landslide monitoring area and the surrounding water area; ( , ) are the coordinates of the boundary point of the seismic landslide monitoring area, that is, the starting point coordinates; ( , ) are the coordinates of the boundary point of the water body, that is, the ending point coordinates.

[0020] In step three, the process of obtaining the initial clustering information of the seismic landslide includes: The initial clustering information of the seismic landslide includes the seismic data, topographic and geomorphic data, and meteorological and hydrological data of the initial clustering center point of the seismic landslide; Set the seismic landslide monitoring data points, divide the seismic landslide monitoring area into several small monitoring areas, and represent each small monitoring area through each seismic landslide monitoring data point; According to the seismic data, topographic and geomorphic data, and meteorological and hydrological data of the seismic landslide monitoring area, obtain the seismic data, topographic and geomorphic data, and meteorological and hydrological data corresponding to the seismic landslide monitoring data points; Integrate the seismic data, topographic and geomorphic data, and meteorological and hydrological data within each seismic landslide monitoring data point to generate a feature vector of the seismic landslide monitoring data point; Through the subtractive clustering algorithm, calculate the density of each seismic landslide monitoring data point, take the seismic landslide monitoring data point with the maximum density as the first clustering center point of the seismic landslide, and update the density of the seismic landslide monitoring data point; Compare the density of the updated seismic landslide monitoring data point with the density of the first clustering center point of the seismic landslide. If the density of the updated seismic landslide monitoring data point is less than the density of the first clustering center point of the seismic landslide, then take the first clustering center point of the seismic landslide as the initial clustering center point of the seismic landslide; if the density of the updated seismic landslide monitoring data point is greater than the density of the first clustering center point of the seismic landslide, then take the updated seismic landslide monitoring data point as the initial clustering center point of the seismic landslide, and repeat the update of the density of the seismic landslide monitoring data point until the seismic landslide monitoring data point with the maximum density is obtained and taken as the initial clustering center point of the seismic landslide; The calculation process of the density of the seismic landslide monitoring data point and the density of the updated seismic landslide monitoring data point includes: Among them, is the density of the seismic landslide monitoring data point; is the density of the updated seismic landslide monitoring data point; and are the feature vectors of the seismic landslide monitoring data point i and the seismic landslide monitoring data point j respectively; and are the influence range parameters; is the feature vector of the h-th clustering center point; is the density value of the h-th clustering center point; Through the feature vector of the seismic landslide monitoring data point, correspondingly obtain the seismic data, topographic and geomorphic data, and meteorological and hydrological data of the initial clustering center point of the seismic landslide.

[0021] In step four, the construction process of the ANFIS model includes: Take the initial clustering information of the seismic landslide as the input variable of the ANFIS model, define the fuzzy membership function and fuzzy set for each input variable, substitute the actual input variable data into the corresponding fuzzy membership function, and convert each input variable into a fuzzy membership value to realize the fuzzification of the input data; Specifically, for earthquake intensity, the corresponding fuzzy sets are defined as low intensity, medium intensity, and high intensity; for earthquake frequency, the corresponding fuzzy sets are defined as low frequency, medium frequency, and high frequency; for earthquake duration, the corresponding fuzzy sets are defined as very short, short, medium, long, and very long; for peak ground acceleration, the corresponding fuzzy sets are defined as very small, small, medium, large, and very large; for epicentral distance, the corresponding fuzzy sets are defined as very close, close, relatively close, far, and very far; for slope, the corresponding fuzzy sets are defined as very gentle, gentle, relatively steep, steep, and very steep; for elevation, the corresponding fuzzy sets are defined as very low, low, medium, high, and very high; for planar curvature, the corresponding fuzzy sets are defined as very concave, concave, nearly horizontal, convex, and very convex; for profile curvature, the corresponding fuzzy sets are defined as very concave, concave, nearly horizontal, convex, and very convex; for air temperature, the corresponding fuzzy sets are defined as very cold, cold, moderate, hot, and very hot; for wind speed, the corresponding fuzzy sets are defined as very small, small, medium, large, and very large; for rainfall, the corresponding fuzzy sets are defined as light rain, moderate rain, heavy rain, and rainstorm; for water level, the corresponding fuzzy sets are defined as very low, relatively low, medium, high, and very high; for water discharge, the corresponding fuzzy sets are defined as very small, small, medium, large, and very large; for sediment concentration, the corresponding fuzzy sets are defined as very little, little, medium, much, and very much; for water velocity, the corresponding fuzzy sets are defined as very slow, slow, medium speed, fast, and very fast; Based on the initial clustering center points of earthquake-induced landslides obtained by the subtractive clustering algorithm, determine the number of fuzzy rules. Through the logical combination between each input variable and the corresponding fuzzy set, using the AND relationship and the OR relationship, construct a fuzzy rule base; According to the fuzzy rule base, perform inference operations on the fuzzy results of the input variables, obtain the output results of each fuzzy rule, and synthesize the outputs of all fuzzy rules to obtain the total fuzzy rule output result; Through the maximum membership degree method, perform defuzzification processing on the total fuzzy rule output result, select the domain value corresponding to the maximum membership degree as the output value of the ANFIS model, and then construct the ANFIS model. Among them, the domain value corresponding to the maximum membership degree is the interval of the fuzzy membership function when the membership degree of the input variable in the fuzzy set reaches the maximum value, and the output value of the ANFIS model is the earthquake-induced landslide hazard index.

[0022] In step four, the process of training the ANFIS model and obtaining the subtractive clustering ANFIS model includes: Extract the earthquake data, topographic and geomorphic data, and meteorological and hydrological data from the earthquake-induced landslide monitoring dataset. Using the training set data, combined with the backpropagation algorithm, take the earthquake data, topographic and geomorphic data, and meteorological and hydrological data as inputs, and the earthquake-induced landslide hazard index as the output, learn the non-linear relationship between the earthquake data, topographic and geomorphic data, meteorological and hydrological data and the earthquake-induced landslide hazard index, and train the ANFIS model; Input the test set data into the ANFIS model, compare the output results of the ANFIS model with the actual seismic landslide hazard index, adjust the parameters of the ANFIS model, optimize the ANFIS model, and obtain the subtractive clustering ANFIS model.

[0023] In step five, the process of obtaining the seismic landslide hazard index for each geographical grid unit includes: According to the small monitoring areas corresponding to the seismic landslide monitoring data points, divide the seismic landslide monitoring area into each geographical grid, and then obtain the seismic landslide monitoring data of each geographical grid unit; Combined with seismic data, topographic and geomorphic data, meteorological and hydrological data, and the subtractive clustering ANFIS model, obtain the seismic landslide hazard index for each geographical grid unit.

[0024] In step six, the process of obtaining the hazard calibration coefficient includes: Extract the feature extraction data in the seismic landslide monitoring data set, the water level, flow rate, sediment concentration, and flow velocity of the water body around the seismic landslide monitoring area, and comprehensively analyze the vegetation coverage, lithology, and soil composition of the monitoring area as input parameters. Integrate the extracted input parameters to construct an input data set; Combined with the input data set and the multiple linear regression algorithm, use the extracted input parameters as input data and the hazard calibration coefficient as output data to learn the linear relationship between the input data and the output data, and train the seismic landslide calibration model; Input the input data set into the seismic landslide calibration model, adjust the intercept term and regression coefficient of the seismic landslide calibration model, optimize the performance of the seismic landslide calibration model, deploy the optimized seismic landslide calibration model into the system, and combine the current input parameters to output the corresponding hazard calibration coefficient; The expression of the seismic landslide calibration model is: Among them, is the hazard calibration coefficient; , , , are the regression coefficients of each input parameter respectively; , , are each input parameter respectively; and are the intercept term and error term of the seismic landslide calibration model respectively.

[0025] In step seven, the process of calibrating the seismic landslide hazard index of each geographical grid unit and dividing the seismic landslide hazard level within each geographical grid unit includes: Calibrate the seismic landslide hazard index of each geographical grid unit using the hazard calibration coefficient to obtain the calibrated seismic landslide hazard index of each geographical grid unit; Set the first threshold of the seismic landslide hazard index and the second threshold of the seismic landslide hazard index. When the seismic landslide hazard index of each geographical grid unit is lower than the first threshold of the seismic landslide hazard index, the corresponding geographical grid unit is of a low seismic landslide hazard level; when the seismic landslide hazard index of each geographical grid unit is between the first threshold and the second threshold of the seismic landslide hazard index, the corresponding geographical grid unit is of a medium seismic landslide hazard level; when the seismic landslide hazard index of each geographical grid unit is higher than the first threshold of the seismic landslide hazard index, the corresponding geographical grid unit is of a high seismic landslide hazard level.

[0026] In Step 8, the process of drawing the seismic landslide hazard zoning map includes: Import the seismic landslide hazard levels within each geographical grid unit into GIS software. Using the GIS software, set green icons, yellow icons, and red icons for the low seismic landslide hazard level, medium seismic landslide hazard level, and high seismic landslide hazard level respectively, and then obtain the seismic landslide hazard level layer; Convert the seismic data, topographic and geomorphic data, and meteorological and hydrological data of each geographical grid unit into raster data, and perform visualization settings on this raster data to obtain the seismic data layer, topographic and geomorphic data layer, and meteorological and hydrological data layer respectively; Overlay the seismic landslide hazard level layer, seismic data layer, topographic and geomorphic data layer, and meteorological and hydrological data layer to draw the seismic landslide hazard zoning map.

[0027] As described above, it is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application. Therefore, the protection scope of this application shall be subject to the protection scope of the claims.

Claims

1. A method for zoning earthquake landslide hazard based on reduced clustering ANFIS, characterized by: The following steps are involved: Step 1: Collect earthquake landslide monitoring data including earthquake data, topographic data, meteorological and hydrological data, and image data; Step 2: extract features from the image data to obtain feature extraction data; Step 3: Based on the earthquake landslide monitoring data, a cluster reduction algorithm is used to determine the initial cluster center point of the earthquake landslide area, and then obtain the initial cluster information of the earthquake landslide; Step 4: Based on the obtained initial clustering information of earthquake landslides, an ANFIS model is constructed, and the ANFIS model is trained to obtain the final cluster reduction ANFIS model; Step 5: Divide the earthquake-landslide monitoring area into various geographic grids, assign corresponding earthquake-landslide monitoring data to each geographic grid unit, and obtain the earthquake-landslide hazard index of each geographic grid unit by combining the cluster reduction ANFIS model; Step 6: Combine meteorological and hydrological data, feature extraction data and neural network algorithm to build an earthquake landslide calibration model, and then obtain the hazard calibration coefficient; Step 7: calibrate the earthquake landslide hazard index of each geographic grid unit using the hazard calibration coefficient, and divide the earthquake landslide hazard level in each geographic grid unit; Step 8. Draw an earthquake landslide hazard zoning map based on the earthquake landslide hazard level and earthquake landslide monitoring data in each geographic grid unit.

2. The earthquake landslide hazard zoning method based on reduced clustering ANFIS according to claim 1 is characterized by: In the step 1, the collection process of earthquake landslide monitoring data includes: Deploy different types of acquisition equipment, combine geological monitoring stations, acquisition equipment and regional databases to collect earthquake data, topographic data, meteorological and hydrological data and image data in the earthquake and landslide monitoring area, wherein the acquisition equipment includes total stations, altitude meters, temperature sensors, ultrasonic anemometers, optical rain gauges, float-type water level gauges, electromagnetic flow meters, sediment content meters, electromagnetic current meters and optical cameras; The earthquake data include earthquake intensity, earthquake frequency, earthquake duration, earthquake peak acceleration and epicenter distance; the topographic data include vegetation coverage, lithology, soil composition, slope, altitude, plane curvature and profile curvature; the meteorological and hydrological data include temperature, wind speed, precipitation, and water level, flow, sand content and flow velocity of surrounding water bodies; the image data are remote sensing images of earthquake landslide monitoring areas; The collected seismic data, topographic data, meteorological and hydrological data are cleaned and standardized, the collected image data are enhanced and denoised, the preprocessed seismic data, topographic data, meteorological and hydrological data are integrated to generate an earthquake landslide monitoring data set, and the earthquake landslide monitoring data set is divided into a training set and a test set.

3. The earthquake landslide hazard zoning method based on reduced clustering ANFIS according to claim 2 is characterized by: In step 2, the process of feature extraction data includes: The remote sensing images of the earthquake landslide monitoring area are imported into ENVI software, and the grayscale values ​​of the green light band and the near infrared band are extracted from the remote sensing images of the earthquake landslide monitoring area by using RNVI software, and then the grayscale values ​​of the green light band and the grayscale values ​​of the near infrared band are converted into the reflectance values ​​of the green light band and the reflectance values ​​of the near infrared band respectively; Based on the reflectance values ​​of the green light band and the near-infrared band, the normalized water index is calculated using the multispectral index method; Set a normalized water index threshold, and identify water areas and non-water areas in the earthquake landslide monitoring area based on the normalized water index threshold; Through GIS software, the coordinates of each point in the earthquake landslide monitoring area are obtained. The boundary point of the earthquake landslide monitoring area is taken as the starting point, and the boundary point of the water body is taken as the end point. Combined with the Euclidean distance formula, the distance from the starting point to the end point is calculated, that is, the distance between the earthquake landslide monitoring area and the surrounding water area, which is then used as feature extraction data and integrated into the earthquake landslide monitoring data set.

4. The earthquake landslide hazard zoning method based on reduced clustering ANFIS according to claim 3 is characterized by: In step 3, the process of obtaining the initial clustering information of earthquake landslides includes: The earthquake landslide initial clustering information includes earthquake data, topographic data, and meteorological and hydrological data of the earthquake landslide initial clustering center point; Set earthquake landslide monitoring data points, divide the earthquake landslide monitoring area into several small monitoring areas, and use each earthquake landslide monitoring data point to represent each small monitoring area; According to the earthquake data, topographic data and meteorological and hydrological data of the earthquake and landslide monitoring area, the earthquake data, topographic data and meteorological and hydrological data of the corresponding earthquake and landslide monitoring data points are obtained; Integrate the earthquake data, topographic data, and meteorological and hydrological data in each earthquake-landslide monitoring data point to generate a feature vector of the earthquake-landslide monitoring data point; The density of each earthquake landslide monitoring data point is calculated by using the cluster reduction algorithm, the earthquake landslide monitoring data point with the largest density is used as the first cluster center point of the earthquake landslide, and the density of the earthquake landslide monitoring data point is updated; Compare the density of the updated earthquake landslide monitoring data points with the density of the first cluster center point of the earthquake landslide. If the density of the updated earthquake landslide monitoring data points is less than the density of the first cluster center point of the earthquake landslide, then use the first cluster center point of the earthquake landslide as the initial cluster center point of the earthquake landslide; if the density of the updated earthquake landslide monitoring data points is greater than the density of the first cluster center point of the earthquake landslide, then use the updated earthquake landslide monitoring data points as the initial cluster center point of the earthquake landslide, and repeatedly update the density of the earthquake landslide monitoring data points until the earthquake landslide monitoring data points with the largest density are obtained, and use them as the initial cluster center point of the earthquake landslide; Through the characteristic vector of earthquake-landslide monitoring data points, the seismic data, topographic data, and meteorological and hydrological data of the initial clustering center point of the earthquake-landslide are obtained.

5. The earthquake landslide hazard zoning method based on reduced clustering ANFIS according to claim 4 is characterized by: In step 4, the construction process of the ANFIS model includes: The initial clustering information of earthquake landslide is used as the input variable of ANFIS model, and the fuzzy membership function and fuzzy set are defined for each input variable. The actual input variable data is substituted into the corresponding fuzzy membership function, and each input variable is converted into a fuzzy membership value to realize the fuzzification of input data. Based on the initial clustering center points of earthquake landslides obtained by the cluster reduction algorithm, the number of fuzzy rules is determined, and the fuzzy rule base is constructed by the logical combination between each input variable and the corresponding fuzzy set, using the AND relationship and the OR relationship; According to the fuzzy rule base, the fuzzification results of the input variables are inferred to obtain the output results of each fuzzy rule, and the outputs of all fuzzy rules are synthesized to obtain the total fuzzy rule output results; The total fuzzy rule output result is defuzzified by the maximum membership method, and the domain value corresponding to the maximum membership is selected as the output value of the ANFIS model, and then the ANFIS model is constructed. The domain value corresponding to the maximum membership is the interval of the fuzzy membership function corresponding to the membership of the input variable in the fuzzy set reaches the maximum value. The output value of the ANFIS model is the earthquake landslide hazard index.

6. The earthquake landslide hazard zoning method based on reduced clustering ANFIS according to claim 5 is characterized by: In step 4, the process of training the ANFIS model and obtaining the reduced clustering ANFIS model includes: Extract earthquake data, topographic data, meteorological and hydrological data from the earthquake landslide monitoring data set, use the training set data, combined with the back propagation algorithm, take earthquake data, topographic data, meteorological and hydrological data as input, take earthquake landslide hazard index as output, learn the nonlinear relationship between earthquake data, topographic data, meteorological and hydrological data and earthquake landslide hazard index, and train the ANFIS model; The test set data is input into the ANFIS model, the output results of the ANFIS model are compared with the actual earthquake landslide hazard index, the ANFIS model parameters are adjusted, the ANFIS model is optimized, and the cluster reduction ANFIS model is obtained.

7. The earthquake landslide hazard zoning method based on reduced clustering ANFIS according to claim 6 is characterized by: In step 5, the process of obtaining the earthquake landslide hazard index of each geographic grid unit includes: According to the small monitoring areas corresponding to the earthquake landslide monitoring data points, the earthquake landslide monitoring area is divided into various geographic grids, and then the earthquake landslide monitoring data of each geographic grid unit is obtained; The earthquake landslide hazard index of each geographic grid unit is obtained by combining earthquake data, topographic data, meteorological and hydrological data with the cluster reduction ANFIS model.

8. The earthquake landslide hazard zoning method based on reduced clustering ANFIS according to claim 7 is characterized by: In step 6, the process of obtaining the risk calibration coefficient includes: Extract the feature extraction data from the earthquake landslide monitoring data set, the water level, flow, sediment content and flow velocity of the water body around the earthquake landslide monitoring area, and comprehensively analyze the vegetation coverage, lithology and soil composition of the monitoring area as input parameters, integrate the extracted input parameters, and construct the input data set; Combining the input data set with the multivariate linear regression algorithm, the extracted input parameters are used as input data, and the hazard calibration coefficient is used as output data to learn the linear relationship between the input data and the output data, and train the earthquake landslide calibration model; The input data set is input into the earthquake-landslide calibration model, the intercept term and regression coefficient of the earthquake-landslide calibration model are adjusted, the performance of the earthquake-landslide calibration model is optimized, the optimized earthquake-landslide calibration model is deployed into the system, and the corresponding hazard calibration coefficient is output in combination with the current input parameters.

9. The earthquake landslide hazard zoning method based on reduced clustering ANFIS according to claim 8 is characterized by: In step 7, the process of calibrating the earthquake landslide hazard index of each geographic grid unit and dividing the earthquake landslide hazard level in each geographic grid unit includes: Using the hazard calibration coefficient, calibrate the earthquake landslide hazard index of each geographic grid unit to obtain the earthquake landslide hazard index of each geographic grid unit after calibration; The first threshold value of the earthquake landslide hazard index and the second threshold value of the earthquake landslide hazard index are set. When the earthquake landslide hazard index of each geographic grid unit is lower than the first threshold value of the earthquake landslide hazard index, the corresponding geographic grid unit is at a low earthquake landslide hazard level; when the earthquake landslide hazard index of each geographic grid unit is between the first threshold value of the earthquake landslide hazard index and the second threshold value of the earthquake landslide hazard index, the corresponding geographic grid unit is at a medium earthquake landslide hazard level; when the earthquake landslide hazard index of each geographic grid unit is higher than the first threshold value of the earthquake landslide hazard index, the corresponding geographic grid unit is at a high earthquake landslide hazard level.

10. The earthquake landslide hazard zoning method based on reduced clustering ANFIS according to claim 9, characterized in that: In step eight, the process of drawing the earthquake landslide hazard zoning map includes: The earthquake landslide hazard level in each geographic grid unit is imported into the GIS software. The GIS software is used to set green icons, yellow icons and red icons for low earthquake landslide hazard level, medium earthquake landslide hazard level and high earthquake landslide hazard level, respectively, thereby obtaining an earthquake landslide hazard level layer. The earthquake data, topographic data and meteorological and hydrological data of each geographic grid unit are converted into raster data, and the raster data is visualized to obtain earthquake data layers, topographic data layers and meteorological and hydrological data layers respectively; The earthquake-landslide hazard level layer, earthquake data layer, topographic data layer and meteorological and hydrological data layer are superimposed to draw an earthquake-landslide hazard zoning map.

Images