Geological disaster identification method and system combined with remote sensing technology

Abstract

This application provides a geological hazard identification method and system combining remote sensing technology. First, a multi-temporal remote sensing image dataset of the target area is acquired, including remote sensing images collected at different times. Then, topographic feature extraction processing is performed on the dataset to obtain a topographic feature set, covering surface cover features, slope features, and elevation change features. Next, multi-temporal change analysis is performed on the topographic feature set to generate a dynamic change feature set. Anomaly area identification is performed on the dynamic change feature set using a geological hazard identification model to determine the set of anomaly areas with geological hazard risk. Finally, a disaster early warning signal is generated based on the anomaly area set and sent to a disaster monitoring terminal. This achieves comprehensive and dynamic monitoring and accurate early warning of geological hazard risk areas, improving the accuracy and efficiency of geological hazard monitoring and early warning.

Description

Technical Field

[0001] This application relates to the field of geological disaster monitoring technology, and more specifically, to a geological disaster identification method and system that combines remote sensing technology. Background Technology

[0002] In the field of geological disaster monitoring and early warning, timely and accurate identification of potential geological disaster risk areas is crucial to safeguarding people's lives and property. Traditional methods of geological disaster identification mainly rely on on-site investigations and geological surveys. While these methods can detect potential geological disaster hazards to some extent, they are often time-consuming and labor-intensive, and limited by the scope of investigation and the experience of personnel, making it difficult to achieve large-scale and efficient monitoring.

[0003] With the rapid development of remote sensing technology, the identification of geological hazards using remote sensing imagery has become a new research hotspot. However, most existing remote sensing geological hazard identification methods are based on single-time-series remote sensing images, making it difficult to capture the dynamic changes before a geological hazard occurs, thus limiting the accuracy and timeliness of identification. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a geological hazard identification method and system that combines remote sensing technology.

[0005] In conjunction with the first aspect of this application, a method for identifying geological hazards using remote sensing technology is provided, applied to a geological hazard identification system using remote sensing technology, the method comprising:

[0006] Obtain a multi-temporal remote sensing image dataset of the target area, wherein the multi-temporal remote sensing image dataset contains remote sensing images acquired at different times;

[0007] The multi-temporal remote sensing image dataset is processed to extract terrain features to obtain a set of terrain features for the target area. The set of terrain features includes land cover features, slope features, and elevation change features.

[0008] The terrain feature set is subjected to multi-temporal change analysis to generate a dynamic change feature set of the target area.

[0009] The geological hazard identification model is invoked to perform abnormal area identification processing on the dynamic change feature set, thereby determining the set of abnormal areas in the target area that pose a risk of geological hazard.

[0010] A disaster early warning signal is generated based on the set of abnormal areas, and the disaster early warning signal is sent to the disaster monitoring terminal.

[0011] In conjunction with a second aspect of this application, a geological hazard identification system incorporating remote sensing technology is provided. The geological hazard identification system incorporating remote sensing technology includes a machine-readable storage medium and a processor. The machine-readable storage medium stores machine-executable instructions. When the processor executes the machine-executable instructions, the geological hazard identification system incorporating remote sensing technology implements the aforementioned geological hazard identification method incorporating remote sensing technology.

[0012] In conjunction with a third aspect of this application, a computer-readable storage medium is provided, wherein computer-executable instructions are stored therein, and when the computer-executable instructions are executed, the aforementioned geological disaster identification method incorporating remote sensing technology is implemented.

[0013] Combining any of the above aspects, by acquiring multi-temporal remote sensing image datasets of the target area and performing terrain feature extraction and multi-temporal change analysis, the dynamic terrain change characteristics of the target area are comprehensively captured. Using a geological hazard identification model to identify anomalous areas based on these dynamic change characteristics, the set of anomalous areas with geological hazard risk in the target area can be accurately and efficiently determined. This not only improves the accuracy and timeliness of geological hazard identification but also overcomes the limitations of traditional methods that are restricted to single-temporal remote sensing images. It achieves comprehensive and dynamic monitoring of geological hazard risk areas. Simultaneously, disaster early warning signals are generated based on the set of anomalous areas and promptly sent to the disaster monitoring terminal to facilitate disaster emergency response and help reduce casualties and property losses caused by geological disasters. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained in conjunction with these drawings without creative effort.

[0015] Figure 1 A flowchart illustrating the geological disaster identification method combining remote sensing technology provided in this application embodiment. Detailed Implementation

[0016] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or end that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or ends.

[0018] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0019] Figure 1 This document illustrates a flowchart of a geological hazard identification method combining remote sensing technology according to an embodiment of this application. It should be understood that in other embodiments, the order of some steps in this geological hazard identification method combining remote sensing technology may be shared based on actual needs, or some steps may be omitted or maintained. The detailed components of this geological hazard identification method combining remote sensing technology are as follows:

[0020] Step S110: Obtain a multi-temporal remote sensing image dataset of the target area, wherein the multi-temporal remote sensing image dataset contains remote sensing images acquired at different times.

[0021] In this embodiment, the primary task for conducting geological hazard identification using remote sensing technology is to acquire a multi-temporal remote sensing image dataset of the target area. This dataset forms the basis for subsequent analysis, containing remote sensing images acquired at different times and reflecting the changing surface information of the target area over time.

[0022] To acquire this dataset, various remote sensing platforms can be used. Satellite remote sensing is one of the commonly used methods; for example, some high-resolution optical satellites can scan and image the Earth's surface according to predetermined orbits and periods. After determining the target area, image data of that area at multiple different time points is collected according to the satellite's revisit period. Assuming the satellite's revisit period is A hours, images at multiple time points separated by multiples of A hours can be collected. These images have the same spatial coverage and can form continuous data in a time series.

[0023] Besides satellite remote sensing, aerial remote sensing is also an effective data acquisition method. Aircraft equipped with remote sensing devices can be used to collect targeted images of specific target areas as needed. The advantage of aerial remote sensing lies in its ability to flexibly adjust the acquisition time and range according to actual conditions, and to acquire higher-resolution image data. When conducting aerial remote sensing data acquisition, it is necessary to rationally plan the flight route and altitude based on factors such as the size and terrain of the target area to ensure that the acquired image data is of good quality and covers the entire target area.

[0024] After collecting remote sensing images from different time periods, these images need to undergo initial screening and processing. The quality of the images is checked, including sharpness and completeness. Images with issues such as cloud cover or missing data need to be processed or removed accordingly. Finally, remote sensing images from different time periods that meet the quality requirements are combined to form a multi-time-series remote sensing image dataset.

[0025] Step S120: Perform terrain feature extraction processing on the multi-time series remote sensing image dataset to obtain a terrain feature set of the target area. The terrain feature set includes land cover features, slope features, and elevation change features.

[0026] After acquiring the multi-time series remote sensing image dataset, the next step is to extract and process its terrain features to obtain a set of terrain features for the target area. This set of terrain features includes land cover features, slope features, and elevation change features. These features are of great significance for the identification of geological hazards and can reflect the terrain conditions of the target area from different perspectives.

[0027] Step S121: Perform radiometric correction on each remote sensing image in the multi-time series remote sensing image dataset to obtain a standardized remote sensing image set.

[0028] Each remote sensing image in a multi-time-series remote sensing image dataset may have radiometric values ​​that are influenced by various factors such as sensor characteristics, atmospheric conditions, and the angle of sunlight. To eliminate these biases, radiometric correction processing is required for each remote sensing image.

[0029] There are various methods for radiometric correction, with calibration coefficient-based methods being common. Assuming the sensor records calibration coefficient B when acquiring an image, and the original radiometric value of the image is C, then the radiometrically corrected radiometric value D can be calculated using the formula D = B × C. By performing this radiometric correction calculation on each image in a multi-temporal remote sensing image dataset, standardized radiometric values ​​can be obtained, thus forming a standardized remote sensing image set. Images processed in this way exhibit consistent radiometric characteristics, facilitating subsequent analysis and processing.

[0030] Step S122: Perform multi-band fusion processing on the standardized remote sensing image set to generate a fused image feature map.

[0031] Each image in a standardized remote sensing image set typically contains information from multiple bands, with different bands reflecting different ground features of the target area. To fully utilize this multi-band information, multi-band fusion processing is required for the standardized remote sensing image set.

[0032] There are many methods for multi-band fusion, such as weighted fusion. Assume a standardized remote sensing image set contains E bands of images, each with weights F1, F2, ..., FE, and corresponding image data G1, G2, ..., GE. In weighted fusion, each band image data is first multiplied by its corresponding weight, and then these results are combined. Specifically, for each pixel in the image, its value in each band is multiplied by its corresponding weight, and then the results are combined in a specific order to generate a fused image feature map. This fused image feature map integrates information from multiple bands, providing a more comprehensive reflection of the ground features of the target area.

[0033] Step S123: Perform land cover classification processing on the fused image feature map to obtain the land cover features of the target area. The land cover features include vegetation coverage, water body distribution range, and bare land area identification.

[0034] Step S1231: Call the pre-trained land cover classification model to perform pixel-level classification processing on the fused image feature map and generate an initial classification result map.

[0035] To obtain the land cover features of the target area, it is necessary to perform land cover classification processing on the fused image feature map. Here, a pre-trained land cover classification model can be called. This land cover classification model is trained with a large amount of labeled sample data and can identify different types of land cover.

[0036] The fused image feature map is input into a pre-trained land cover classification model, which then classifies each pixel in the image. Based on the pixel's spectral and texture features, the model determines which land cover type—vegetation, water, bare land, etc.—the pixel belongs to. After processing, the model generates an initial classification result map, in which each pixel is assigned a label for a land cover type.

[0037] Step S1232: Perform spatial smoothing on the initial classification result map, and perform regional merging on the spatially smoothed classification result map to generate continuous land cover area identifiers.

[0038] The initial classification result image may contain some noise and discontinuous classification regions. To make the classification results more accurate and consistent, spatial smoothing is required. Spatial smoothing can be achieved using filtering algorithms, such as mean filtering. For each pixel in the initial classification result image, a neighborhood of a certain size is selected centered on that pixel. The mean of the land cover type for all pixels within that neighborhood is calculated, and this mean is used as the new land cover type for that pixel. This process reduces the influence of noise, making the classification results smoother.

[0039] After spatial smoothing, the classification results map also needs to undergo region merging. Some adjacent small regions may belong to the same land cover type, but they are labeled separately due to classification errors or other reasons. Through region merging, these adjacent small regions belonging to the same land cover type are merged into a continuous large region, thereby generating continuous land cover area labels.

[0040] Step S1233: Calculate the area percentage corresponding to each land cover area identifier, generate vegetation coverage, water body distribution range and bare land area identifier, and obtain the land cover characteristics of the target area.

[0041] After obtaining continuous land cover area identifiers, it is necessary to calculate the area percentage corresponding to each land cover area identifier. For each land cover type, calculate the proportion of its corresponding area within the entire target area. For example, for vegetated areas, calculate its area percentage; this proportion represents the vegetation cover degree. For water body distribution areas, calculate its area percentage to obtain information about the water body distribution range. For bare land areas, similarly calculate its area percentage and identify the area. Through this statistical analysis and identification, the land cover characteristics of the target area are obtained, including vegetation cover degree, water body distribution range, and bare land area identifiers.

[0042] Step S124: Perform digital elevation model reconstruction processing on the fused image feature map to generate an elevation data matrix.

[0043] Digital elevation models (DEMs) can reflect the topographic relief of a target area and are crucial for extracting slope and elevation change features in geological hazard identification. To generate a digital elevation model of the target area, it is necessary to perform digital elevation model reconstruction processing on the feature map of the fused image.

[0044] There are many methods for digital elevation model (DEM) reconstruction, such as the stereo pair matching method. Suitable stereo pairs are selected from the feature map of the fused image, and the disparity of these points is calculated by matching corresponding points within the stereo pairs. Based on the disparity information and photogrammetry principles, the elevation value corresponding to each corresponding point can be calculated. Arranging these elevation values ​​according to the pixel positions in the image generates an elevation data matrix, where each element corresponds to the elevation value of a pixel in the target area.

[0045] Step S125: Calculate the slope change gradient based on the elevation data matrix to obtain the slope characteristics of the target area.

[0046] After obtaining the elevation data matrix, the slope gradient can be calculated based on this matrix, thereby obtaining the slope characteristics of the target area. The slope gradient reflects the degree of inclination and changes in the terrain.

[0047] The slope gradient can be calculated using the finite difference method. For each pixel in the elevation data matrix, its neighboring pixels are selected, and the slope gradient of that pixel is obtained by calculating the elevation difference and horizontal distance between the neighboring pixels. For example, for pixel H, its neighboring pixels I and J are selected, and the elevation differences K1 and K2 between H and I, and the horizontal distances L1 and L2 are calculated. Based on these differences and distances, the slope gradient of that pixel in different directions can be calculated. After calculating the slope gradients of all pixels, the slope characteristics of the target area are obtained.

[0048] Step S126: Perform difference calculation on the elevation data matrix corresponding to remote sensing images acquired at different times in the multi-time series remote sensing image dataset to obtain the elevation change characteristics of the target area.

[0049] In a multi-temporal remote sensing image dataset, remote sensing images acquired at different times can all be reconstructed using the aforementioned digital elevation model to obtain corresponding elevation data matrices. To obtain the elevation variation characteristics of the target area, it is necessary to perform interpolation calculations on these elevation data matrices from different times.

[0050] Suppose that the elevation data matrices corresponding to the remote sensing images acquired at times M1 and M2 are N1 and N2, respectively. For each corresponding pixel in the elevation data matrix, calculate the elevation difference between the two time points. For example, for pixel P, its elevation value is Q1 in N1 and Q2 in N2, then the elevation difference R = Q2 - Q1. After calculating the elevation differences for all pixels, the elevation change characteristics of the target area between these two time points are obtained. By performing this difference calculation on the elevation data matrices of different time combinations in a multi-time series remote sensing image dataset, the elevation change characteristics of the target area over multiple time periods can be obtained.

[0051] Step S130: Perform multi-temporal change analysis on the terrain feature set to generate a dynamic change feature set of the target area.

[0052] After obtaining the set of terrain features for the target area, the next step is to perform multi-temporal change analysis on this set to generate a dynamic change feature set for the target area. By analyzing the changes in terrain features over different times, potential geological hazards in the target area can be identified.

[0053] Step S131: Perform time series change detection processing on the land cover features to extract the vegetation cover change rate, water body expansion and contraction trend and bare land migration path.

[0054] Vegetation cover, water body distribution, and bare land area identification in land cover features may change over time. To analyze these changes, time-series change detection processing of land cover features is required.

[0055] For vegetation cover, its rate of change is calculated by comparing the vegetation cover values ​​at different times. Assuming the vegetation cover is T1 at time S1 and T2 at time S2, then the rate of change of vegetation cover U = (T2 - T1) / T1. By performing this calculation on the vegetation cover at multiple time points, a sequence of the rate of change of vegetation cover over time can be obtained, thus allowing analysis of the changing trend of vegetation cover.

[0056] To determine the extent of water body distribution, observe the distribution area at different times to judge whether it is expanding or contracting. The expansion or contraction trend can be determined by comparing the size of the water body distribution area at different times. For example, if the area of ​​the water body distribution area increases within a certain period, it indicates that the water body is expanding; conversely, it indicates contraction.

[0057] For bare land areas, migration paths can be extracted by analyzing changes in their location and shape over different times. Spatial analysis methods can be used to compare features such as the centroid position and boundary shape of bare land areas at different times to determine the migration direction and distance.

[0058] Step S132: Perform time series overlay analysis on the slope features to extract areas of abrupt slope gradient changes and areas of continuous erosion.

[0059] Slope characteristics may change over time. In order to detect anomalies in these changes, it is necessary to perform time series overlay analysis on the slope characteristics.

[0060] Slope feature maps from different times are overlaid to compare the slope changes of each pixel at different times. Areas with significant slope changes, i.e., abrupt slope gradient changes, are marked and extracted. These areas may be due to sudden topographic changes caused by geological activities, human engineering activities, etc., and may pose a risk of geological disasters.

[0061] Simultaneously, analyzing the changing trends of slope characteristics at multiple time points identifies areas where the slope continuously increases, i.e., areas of continuous erosion. These areas, likely due to water erosion, weathering, and other factors leading to gradual topographical erosion, also require close attention.

[0062] Step S133: Perform time series difference processing on the elevation change characteristics to extract the elevation anomaly subsidence area and uplift area.

[0063] Step S1331: Arrange the elevation data matrices from different times in chronological order to generate an elevation time series dataset.

[0064] To analyze the characteristics of elevation changes, the elevation data matrices from different times need to be arranged in chronological order to form an elevation time series dataset. This allows for a clear observation of how the elevation of the target area changes over time. The elevation data matrices for each time point are arranged sequentially according to the order of data collection, with each matrix corresponding to a specific time point.

[0065] Step S1332: Perform pixel-by-pixel difference calculation on the elevation data matrix of adjacent time points in the elevation time series dataset to generate an elevation change rate map.

[0066] After obtaining the elevation time series dataset, pixel-by-pixel difference calculations are performed on the elevation data matrices of adjacent time points. For two adjacent elevation data matrices in the elevation time series dataset, such as elevation data matrices W1 and W2 corresponding to time V1 and time V2, the elevation difference is calculated for each corresponding pixel. Assuming that the elevation value of pixel X in W1 is Y1 and its elevation value in W2 is Y2, then the elevation difference Z for that pixel is Z = Y2 - Y1. After calculating the elevation differences for all pixels, the elevation changes between adjacent time points are obtained. By performing this calculation on multiple adjacent time points, an elevation change rate map can be generated, which reflects the elevation change rate of each pixel in the target area over different time periods.

[0067] Step S1333: Perform threshold segmentation on the elevation change rate map and mark pixels with a change rate exceeding a preset threshold as abnormal change pixels.

[0068] To identify areas of abnormal elevation changes from an elevation change rate map, threshold segmentation is required. A preset elevation change rate threshold is established; for each pixel in the elevation change rate map, if its elevation change rate exceeds this threshold, the pixel is marked as an abnormal pixel. The threshold setting needs to be comprehensively considered based on factors such as the geological background and historical elevation changes of the target area. Threshold segmentation allows for the rapid filtering of pixels with abnormal elevation changes.

[0069] Step S1334: Perform spatial clustering processing on the abnormally changing pixels to generate the spatial boundaries of the elevation anomaly subsidence area and the uplift area.

[0070] After identifying pixels with anomalous changes, spatial clustering is required to determine the specific extent of areas of abnormal elevation subsidence and uplift. This spatial clustering can employ density-based clustering algorithms, such as the DBSCAN algorithm. The DBSCAN algorithm clusters adjacent pixels with higher spatial density into a single class based on their spatial distribution density. Through clustering, the anomalous pixels are divided into different regions, each representing an area of ​​abnormal elevation subsidence or uplift. Then, the spatial boundaries of these regions are generated based on the clustering results, clarifying their specific extent.

[0071] Step S134: Combine the vegetation coverage change rate, water body expansion and contraction trend, bare land migration path, slope gradient abrupt change area, continuous erosion area, elevation abnormal subsidence area and uplift area into a dynamic change feature set.

[0072] After extracting the vegetation cover change rate, water body expansion and contraction trends, bare land migration paths, abrupt slope gradient change areas, areas of continuous erosion, areas of abnormal elevation subsidence, and areas of elevation uplift, these features are merged to form a dynamic change feature set for the target area. This dynamic change feature set integrates dynamic change information of the target area in multiple aspects. These features can be combined according to certain rules, such as arranging them according to feature type or spatial location, to form a unified dynamic change feature set.

[0073] Step S140: Call the geological hazard identification model to perform abnormal area identification processing on the dynamic change feature set, and determine the set of abnormal areas in the target area that have geological hazard risks.

[0074] After obtaining the set of dynamic change characteristics of the target area, the next step is to use a geological hazard identification model to identify anomalous areas within this set, thus determining the set of anomalous areas in the target area that pose a risk of geological hazards. The geological hazard identification model can comprehensively analyze various information from the set of dynamic change characteristics to identify areas that may pose a risk of geological hazards.

[0075] Step S141: Input the dynamic change feature set into the pre-trained geological hazard identification model, which includes a feature fusion module and a risk prediction module.

[0076] The pre-trained geological hazard identification model consists of a feature fusion module and a risk prediction module. When a dynamically changing feature set is input into the model, it first enters the feature fusion module. The function of the feature fusion module is to integrate and optimize various features from the dynamically changing feature set to better provide input for subsequent risk prediction.

[0077] Step S142: The feature fusion module performs weight allocation and feature fusion on the dynamically changing feature set to generate a comprehensive risk feature map.

[0078] Step S1421: Assign dynamic weight coefficients to each dynamic feature in the set of dynamic change features.

[0079] In the feature fusion module, the first step is to assign dynamic weight coefficients to each dynamic feature in the dynamic feature set. Different dynamic features may have varying degrees of impact on geological hazard risk, therefore, different weights need to be assigned based on their importance. For example, areas of abnormal elevation subsidence and areas of abrupt slope gradient changes may have a greater impact on geological hazard risk, so they can be assigned larger weights; while features such as vegetation cover change rate have a relatively smaller impact and can be assigned smaller weights. These weight coefficients are not fixed but can be dynamically adjusted according to the actual situation to adapt to different geological environments and hazard types.

[0080] Step S1422: Based on the dynamic weight coefficients, the various dynamically changing features are weighted and spliced ​​to obtain a preliminary fused feature map.

[0081] After assigning dynamic weight coefficients to each dynamic change feature, the features are weighted and concatenated based on these coefficients. For each dynamic change feature, its data is multiplied by the corresponding weight coefficient, and then these weighted feature data are concatenated in a specific order. For example, features such as vegetation cover change rate and water body expansion / contraction trend are multiplied by their respective weight coefficients and then concatenated to form a preliminary fused feature map. This preliminary fused feature map integrates information from various dynamic change features, but it may still have some shortcomings.

[0082] Step S1423: Perform spatial context enhancement processing on the preliminary fused feature map, and perform residual connection between the spatial context enhanced feature map and the preliminary fused feature map to generate a comprehensive risk feature map.

[0083] Spatial context enhancement aims to capture the spatial relationships between pixels in the initial fused feature map to improve feature representation. One feasible approach is to use convolution operations. Assuming the initial fused feature map is A and the convolution kernel is B, the convolution operation can be viewed as a weighted summation of each pixel and its neighborhood in the initial fused feature map. For each pixel in the initial fused feature map A, a neighborhood of a specific size is selected centered on that pixel. The pixel value within this neighborhood is multiplied by the weight value at the corresponding position in the convolution kernel B. These products are then summed to obtain the new value of the pixel after convolution. By sliding the convolution kernel B across the entire initial fused feature map and performing this convolution calculation on each pixel, a spatial context-enhanced feature map C is generated.

[0084] The size of the convolution kernel B and its weights are key factors affecting the spatial context enhancement effect. The size of the convolution kernel determines the range of the neighborhood; a larger kernel can capture a wider range of spatial information, but the computational cost will increase accordingly; a smaller kernel focuses on extracting local information. The weights are set according to different task requirements and data characteristics, with the aim of highlighting spatial features related to geological hazard risk.

[0085] After obtaining the spatial context-enhanced feature map C, it is residually connected to the initially fused feature map A. The purpose of residual connection is to avoid information loss during feature fusion and ensure that the model can learn subtle changes in the features. Specifically, the pixel values ​​at corresponding positions in feature map C and feature map A are added together to obtain a new feature map D. That is, for pixels at the same position in feature map C and feature map A, their pixel values ​​are added together to obtain the pixel value at the corresponding position in feature map D.

[0086] When performing residual joins, it is necessary to ensure that the dimensions (number of elements) of feature map C and feature map A are matched. If their dimensions are inconsistent, dimensionality adjustment operations may be required, such as interpolation or pooling, to make their dimensions consistent, ensuring the smooth execution of the addition operation. The feature map D obtained after residual joins is the comprehensive risk feature map, which integrates the information from the preliminary fused feature map and the information after spatial context enhancement, and can more comprehensively and accurately reflect the geological hazard risk characteristics of the target area.

[0087] Step S143: The risk prediction module performs binary classification on the comprehensive risk feature map and outputs the geological disaster risk probability of each pixel.

[0088] The risk prediction module receives the comprehensive risk feature map D as input. Its main task is to perform binary classification on each pixel in the comprehensive risk feature map, determine whether there is a geological disaster risk in the area corresponding to the pixel, and output the corresponding risk probability.

[0089] The risk prediction module can employ a neural network model, such as a fully connected neural network. The feature vector of each pixel in the comprehensive risk feature map D is used as input to the fully connected neural network. The fully connected neural network consists of multiple neuron layers, each containing multiple neurons, with neurons in adjacent layers connected by weights. During the forward propagation of the neural network, the input feature vector is processed sequentially through each neuron layer. Each neuron performs a weighted summation of the input signal and a non-linear transformation using an activation function, passing the transformed result to the next layer of neurons.

[0090] Suppose a fully connected neural network has E neurons, with the input feature vector of the input layer being F, the weight matrix of the i-th neuron being Gi, the bias vector being Hi, and the activation function being f. In the first layer, the input feature vector F is multiplied by the weight matrix G1, and the bias vector H1 is added. Then, the result is transformed by the activation function f to obtain the output I1 of the first layer. That is, I1 = f(F*G1 + H1). Next, the output I1 of the first layer is used as the input of the second layer, and the above process is repeated to obtain the output I2 of the second layer, and so on, until the output of the last layer is obtained.

[0091] The output of the final layer is a vector containing two elements, representing the scores for the pixel belonging to the "no geological hazard risk" class and the "geological hazard risk" class, respectively. To convert these scores into probability values, the Softmax function can be used. The Softmax function normalizes the output of the final layer so that the sum of the two elements is 1, and each element represents the probability that the pixel belongs to the corresponding class. For example, if the output vector is (J1, J2), after processing by the Softmax function, we get the probability vector (K1, K2), where K1 represents the probability that the pixel belongs to the "no geological hazard risk" class, K2 represents the probability that the pixel belongs to the "geological hazard risk" class, and K1 + K2 = 1.

[0092] By performing this process on each pixel in the comprehensive risk feature map D, the probability of geological disaster risk for each pixel can be obtained.

[0093] Step S144: Perform regional aggregation processing on the geological disaster risk probability to generate the abnormal region set.

[0094] After obtaining the probability of geological disaster risk for each pixel, these probabilities need to be aggregated regionally to determine the set of anomalous areas where there is a risk of geological disaster.

[0095] Regional aggregation processing can employ a threshold-based approach. First, a geological hazard risk probability threshold L is set. For each pixel in the comprehensive risk feature map, if its geological hazard risk probability is greater than the threshold L, the pixel is marked as a suspected anomalous pixel.

[0096] Next, connectivity analysis is performed on these suspected anomalous pixels. Connectivity component labeling algorithms, such as 4-connectivity or 8-connectivity algorithms, are used to divide adjacent suspected anomalous pixels into different connected regions. The 4-connectivity algorithm considers a pixel to be connected to its four adjacent pixels (top, bottom, left, and right); the 8-connectivity algorithm considers a pixel to be connected to its eight adjacent pixels (top, bottom, left, right, top-left, top-right, bottom-left, and bottom-right).

[0097] For each connected region, statistical measures such as the sum or average of the geological hazard risk probabilities of all pixels within that region are calculated to assess the overall geological hazard risk level of the region. Based on these statistics, connected regions with higher risk levels are selected and identified as anomalous regions.

[0098] Finally, all identified anomalous areas are integrated to generate an anomalous area set. This set contains information on all anomalous areas in the target area that pose a geological hazard risk, such as their location and extent.

[0099] Step S150: Generate a disaster early warning signal based on the set of abnormal areas, and send the disaster early warning signal to the disaster monitoring terminal.

[0100] After identifying the set of anomalous areas in the target region that pose a risk of geological disasters, it is necessary to generate a disaster early warning signal based on this set and send the signal to the disaster monitoring terminal so that timely countermeasures can be taken.

[0101] Step S151: Perform risk level classification on each abnormal region in the abnormal region set to generate high-risk, medium-risk, and low-risk region identifiers.

[0102] Step S1511: Obtain the geological disaster risk probability, historical disaster frequency, and surrounding population density data for each abnormal area.

[0103] In order to accurately classify the risk level of abnormal areas, it is necessary to obtain data on the probability of geological disaster risk, the frequency of historical disasters, and the surrounding population density for each abnormal area.

[0104] The probability of geological disaster risk has been obtained in step S143, that is, the statistical value of the probability of geological disaster risk of all pixels in each abnormal area, such as the average or the sum.

[0105] Historical disaster frequency data can be obtained by consulting historical geological disaster records. For each anomalous area, the number of geological disasters that occurred in that area over a past period is counted, and then the historical disaster frequency is calculated based on the time span.

[0106] Surrounding population density data can be obtained through census data, geographic information system (GIS) data, and other channels. For each anomalous area, the population within a certain surrounding range is determined, and the surrounding population density is calculated based on the area of ​​that range.

[0107] Step S1512: Input the geological disaster risk probability, historical disaster frequency and surrounding population density data into the risk level assessment model, and output the risk score for each abnormal area.

[0108] The risk level assessment model is a trained model that can output the risk score of each abnormal area based on the input geological disaster risk probability, historical disaster occurrence frequency, and surrounding population density data.

[0109] This risk level assessment model can adopt a linear regression model or other machine learning models. Assuming a linear regression model is adopted, its expression is risk score M = a * geological disaster risk probability + b * historical disaster occurrence frequency + c * surrounding population density + d, where a, b, and c are weight coefficients. The geological disaster risk probability, historical disaster occurrence frequency, and surrounding population density data need to be normalized before being input into the risk level assessment model, and d is the bias term. These weight coefficients and bias terms are obtained through learning with a large amount of sample data during the model training process.

[0110] After normalizing and transforming the geological disaster risk probability, historical disaster occurrence frequency, and surrounding population density data of each abnormal area, substitute them into the above expression to calculate the risk score of each abnormal area.

[0111] Step S1513: Divide the risk score into high-risk intervals, medium-risk intervals, and low-risk intervals according to a preset risk score threshold.

[0112] Preset three risk score thresholds N1, N2 (N1 < N2), and divide the risk score into three intervals according to these two thresholds. If the risk score of a certain abnormal area is greater than N2, then this area belongs to the high-risk interval; if the risk score is between N1 and N2, then this area belongs to the medium-risk interval; if the risk score is less than N1, then this area belongs to the low-risk interval.

[0113] Step S1514: Assign corresponding risk level identifiers to each abnormal area.

[0114] According to the risk interval to which the abnormal area belongs, assign corresponding risk level identifiers to each abnormal area. For example, for an abnormal area belonging to the high-risk interval, assign a high-risk area identifier; for an abnormal area belonging to the medium-risk interval, assign a medium-risk area identifier; for an abnormal area belonging to the low-risk interval, assign a low-risk area identifier. …

[0115] Step S152: Generate corresponding warning level signals according to the risk level division result.

[0116] Based on the risk level identifier of each abnormal area, a corresponding early warning level signal is generated. For example, a high-risk area generates a high-level early warning signal; a medium-risk area generates a medium-level early warning signal; and a low-risk area generates a low-level early warning signal. These early warning level signals can be represented by different colors, symbols, or text for clear display on the disaster monitoring terminal.

[0117] Step S153: Associate the warning level signal with the geographical coordinates of the abnormal area to generate a disaster warning signal.

[0118] To enable disaster monitoring terminals to accurately locate abnormal areas, it is necessary to associate the warning level signal with the geographic coordinates of the abnormal area. For each abnormal area, its geographic coordinate information, such as latitude and longitude coordinates, is obtained. Then, the warning level signal for that abnormal area is combined with the geographic coordinate information to generate a disaster warning signal that includes the warning level and geographic coordinates.

[0119] Step S154: Transmit the disaster early warning signal to the disaster monitoring terminal in real time via the wireless communication network to trigger the terminal's early warning prompt.

[0120] Wireless communication networks, such as 4G and 5G, are used to transmit the generated disaster early warning signals to the disaster monitoring terminal in real time. Upon receiving the disaster early warning signal, the disaster monitoring terminal will trigger corresponding alerts, such as sound alarms, vibration alerts, and screen pop-ups, to remind relevant personnel to pay attention to the geological disaster risk in the target area.

[0121] Step S210: Collect remote sensing image datasets of historical geological disaster cases and corresponding disaster annotation data to construct a training sample set.

[0122] To train a geological disaster identification model, it is necessary to collect remote sensing image datasets of historical geological disaster cases and corresponding disaster annotation data to construct a training sample set.

[0123] Remote sensing image datasets of historical geological disaster cases can be obtained from multiple sources, such as satellite image databases and airborne remote sensing data centers. Collecting remote sensing images of geological disasters occurring at different times and in different regions ensures the diversity and representativeness of the images.

[0124] The corresponding disaster labeling data refers to the annotation information of geological disaster areas in remote sensing images. Geological disaster areas in remote sensing images can be labeled through manual visual interpretation or other automated annotation methods, such as the boundaries and extent of disaster areas like landslides and debris flows.

[0125] The collected remote sensing image datasets and corresponding disaster annotation data are organized and paired, with each pair of remote sensing images and disaster annotation data constituting a training sample. All training samples are combined to construct a training sample set.

[0126] Step S220: Perform terrain feature extraction and dynamic change analysis on the remote sensing image dataset in the training sample set to generate a set of sample dynamic change features.

[0127] The process of extracting terrain features and performing dynamic change analysis on the remote sensing image dataset in the training sample set is similar to steps S120-S130.

[0128] First, radiometric correction is performed on the remote sensing image dataset to obtain a standardized remote sensing image set. Then, multi-band fusion processing is performed on the standardized remote sensing image set to generate a fused image feature map. Next, land cover classification processing and digital elevation model reconstruction processing are performed on the fused image feature map to obtain land cover features, slope features, and elevation change features.

[0129] Subsequently, these topographic features were subjected to multi-temporal change analysis, including time series change detection processing of land cover features, time series overlay analysis processing of slope features, and time series difference processing of elevation change features. Dynamic change features such as vegetation cover change rate, water body expansion and contraction trend, bare land migration path, slope gradient abrupt change area, continuous erosion area, elevation abnormal subsidence area and uplift area were extracted.

[0130] Finally, these dynamic change features are merged into a sample dynamic change feature set.

[0131] Step S230: Input the set of dynamic change features of the sample into the initial geological hazard identification model and output the predicted hazard area.

[0132] The generated set of dynamic change features of the samples is input into the initial geological hazard identification model. The structure of the initial geological hazard identification model is the same as that of the pre-trained geological hazard identification model used in step S140, including a feature fusion module and a risk prediction module.

[0133] The feature fusion module performs weight allocation and feature fusion on the dynamically changing feature set of the samples to generate a comprehensive risk feature map. The risk prediction module performs binary classification on the comprehensive risk feature map, outputting the geological disaster risk probability for each pixel. Based on these probabilities, the predicted disaster areas are determined through regional aggregation.

[0134] Step S240: Calculate the difference loss between the predicted disaster area and the disaster labeling data, and update the model parameters based on the difference loss.

[0135] To evaluate the performance of the initial geological hazard identification model, it is necessary to calculate the difference loss between the predicted hazard area and the hazard labeling data. The cross-entropy loss function can be used to calculate the difference loss.

[0136] For each training sample, the predicted geological disaster risk probability of each pixel in the disaster area is compared with the true label (disaster present or no disaster) of the corresponding pixel in the disaster labeling data. The cross-entropy loss function calculates a loss value based on the difference between the two, which reflects the degree of deviation between the model's prediction and the actual situation.

[0137] The sum of the variance losses for all training samples is calculated to obtain the total variance loss. Based on this total variance loss, an optimization algorithm, such as stochastic gradient descent, is used to update the parameters of the initial geological hazard identification model. The stochastic gradient descent algorithm adjusts the values ​​of the model parameters according to the gradient of the variance loss with respect to the model parameters, so that the variance loss gradually decreases.

[0138] Step S250: Repeat the training until the model converges to obtain the pre-trained geological hazard identification model.

[0139] Repeat steps S230-S240, continuously inputting the dynamically changing feature set of the samples into the model, calculating the difference loss, and updating the model parameters. As training progresses, the model's performance will gradually improve, and the difference loss will gradually decrease.

[0140] The model is considered converged when the difference loss no longer decreases significantly or when the preset number of training epochs is reached. At this point, the obtained model is the pre-trained geological hazard identification model, which can be used for actual geological hazard identification tasks.

[0141] Through the above steps, geological disaster identification can be achieved by combining remote sensing technology. Starting from acquiring multi-time-series remote sensing image datasets, the process involves terrain feature extraction, dynamic change analysis, and abnormal area identification, ultimately generating a disaster early warning signal and sending it to the disaster monitoring terminal. It also includes the training process of the geological disaster identification model to ensure that the model can accurately identify geological disaster risk areas.

[0142] In the above embodiments, the geological disaster identification system combining remote sensing technology for performing the above method embodiments has at least one processor, a control module (chipset) coupled to at least one of the processors, a memory coupled to the control module, a non-volatile memory (NVM) / storage device coupled to the control module, at least one load to / output device coupled to the control module, and a network interface coupled to the control module.

[0143] The processor may include at least one single-core or multi-core processor, and may include any combination of general-purpose processors or special-purpose processors (e.g., graphics processors, application processors, baseband processors, etc.). For some alternative implementations, a geological disaster identification system incorporating remote sensing technology can serve as an electronic device such as the gateway described in the embodiments of this application.

[0144] In some alternative implementations, a geological hazard identification system incorporating remote sensing technology may include at least one computer-readable medium (e.g., a memory or NVM / storage device) having instructions and at least one processor fused with the at least one computer-readable medium and configured to execute the instructions to implement the module thereby performing the actions described in this disclosure.

[0145] In one embodiment, the control module may include any suitable interface controller to provide any suitable interface to at least one of the processors and / or any suitable device or component communicating with the control module.

[0146] The control module may include a memory controller module to provide an interface to the memory. The memory controller module may be a hardware module, a software module, and / or a firmware module.

[0147] The memory can be used, for example, to load and store data and / or instructions for a geological hazard identification system incorporating remote sensing technology. In one embodiment, the memory may include any suitable volatile memory, such as suitable DRAM.

[0148] In one embodiment, the control module may include at least one load-to-output controller to provide an interface to the NVM / storage device and (at least one) load-to-output device.

[0149] For example, an NVM / storage device can be used to store data and / or instructions. An NVM / storage device may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (at least one) non-volatile storage device (e.g., at least one hard disk drive (HDD), at least one optical disc (CD) drive, and / or at least one digital universal optical disc (DVD) drive).

[0150] NVM / storage devices may include storage resources that are physically installed as part of a geological hazard identification system incorporating remote sensing technology, or that can be accessed by the device without needing to be part of it. For example, NVM / storage devices may be accessed over a network via (at least one) load-to-output device.

[0151] At least one loading / output device may provide an interface for the geological hazard identification system incorporating remote sensing technology to communicate with any other suitable device. The loading / output device may include communication components, pinyin components, sensor components, etc. A network interface may provide an interface for the geological hazard identification system incorporating remote sensing technology to communicate via at least one network. The geological hazard identification system incorporating remote sensing technology may wirelessly communicate with at least one component of a wireless network based on at least one wireless network prior and / or protocol, such as accessing a wireless network based on communication priors.

[0152] In one embodiment, at least one of the processors may be integrated with the logic of at least one controller of the control module (e.g., a memory controller module). In one embodiment, at least one of the processors may be integrated with the logic of at least one controller of the control module to form a system-level integration. In one embodiment, at least one of the processors may be fused with the logic of at least one controller of the control module on the same die. In one embodiment, at least one of the processors may be fused with the logic of at least one controller of the control module on the same die to form a system-on-a-chip (SoC).

[0153] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0154] This invention discloses a computer read storage medium that stores a computer program for electronic data exchange, wherein the computer program causes a computer to execute the steps in the geological disaster identification method combining remote sensing technology described in the foregoing embodiments.

[0155] This invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform the steps in the geological hazard identification method combining remote sensing technology described in the foregoing embodiments.

[0156] The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any inventive effort.

[0157] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electronically erasable rewritable read-only memory (EEPROM), compact optical disc (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium that can be used to have or store data.

[0158] Finally, it should be noted that the above-disclosed embodiments are merely preferred embodiments of the present invention and are only used to illustrate the technical solutions of the present invention, not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying geological hazards using remote sensing technology, characterized in that, The method includes: Acquire a multi-temporal remote sensing image dataset of the target area, wherein the multi-temporal remote sensing image dataset contains remote sensing images acquired at different times; The terrain feature extraction process is performed on the multi-temporal remote sensing image dataset to obtain a set of terrain features for the target area, including: Each remote sensing image in the multi-time-series remote sensing image dataset is subjected to radiometric correction to obtain a standardized remote sensing image set. The standardized remote sensing image set is subjected to multi-band fusion processing to generate a fused image feature map; The fused image feature map is subjected to land cover classification processing to obtain the land cover features of the target area, which include vegetation coverage, water body distribution range and bare land area identification. The fused image feature map is subjected to digital elevation model reconstruction processing to generate an elevation data matrix; The slope gradient is calculated based on the elevation data matrix to obtain the slope characteristics of the target area; The elevation change characteristics of the target area are obtained by performing difference calculation on the elevation data matrix corresponding to remote sensing images acquired at different times in the multi-time series remote sensing image dataset. The set of terrain features includes land cover features, slope features, and elevation variation features; The terrain feature set is subjected to multi-temporal change analysis to generate a dynamic change feature set of the target area, including: The land cover features are subjected to time series change detection processing to extract vegetation cover change rate, water body expansion and contraction trend and bare land migration path; The slope features were subjected to time series overlay analysis to extract areas of abrupt slope gradient changes and areas of continuous erosion. The elevation change characteristics are subjected to time series differencing to extract abnormal elevation subsidence areas and uplift areas, including: Arrange the elevation data matrices from different times in chronological order to generate an elevation time series dataset. Pixel-by-pixel difference calculation is performed on the elevation data matrix of adjacent time points in the elevation time series dataset to generate an elevation change rate map. The elevation change rate map is subjected to threshold segmentation processing, and pixels with a change rate exceeding a preset threshold are marked as abnormal change pixels; Spatial clustering is performed on the abnormally changing pixels to generate the spatial boundaries of the elevation anomaly subsidence area and the uplift area. The vegetation coverage change rate, the expansion and contraction trend of water body range, the migration path of bare land area, the slope gradient abrupt change area, the continuous erosion area, the elevation abnormal subsidence area and the uplift area are combined into a dynamic change feature set. The geological hazard identification model is invoked to perform anomaly region identification processing on the dynamic change feature set, thereby determining the set of anomaly regions in the target area where there is a risk of geological hazard, including: The dynamic change feature set is input into a pre-trained geological hazard identification model, which includes a feature fusion module and a risk prediction module. The feature fusion module performs weight allocation and feature fusion on the dynamically changing feature set to generate a comprehensive risk feature map. The risk prediction module performs binary classification on the comprehensive risk feature map and outputs the geological disaster risk probability of each pixel. The geological hazard risk probabilities are aggregated regionally to generate the set of abnormal regions; A disaster early warning signal is generated based on the set of abnormal areas, and the disaster early warning signal is sent to the disaster monitoring terminal.

2. The geological hazard identification method combining remote sensing technology according to claim 1, characterized in that, The step of performing land cover classification processing on the fused image feature map to obtain the land cover features of the target area includes: The pre-trained land cover classification model is invoked to perform pixel-level classification processing on the fused image feature map to generate an initial classification result map; The initial classification result map is spatially smoothed, and the spatially smoothed classification result map is then merged to generate continuous land cover area identifiers. The area percentage corresponding to each land cover area is statistically analyzed to generate vegetation coverage, water body distribution range, and bare land area identifiers, thereby obtaining the land cover characteristics of the target area.

3. The geological hazard identification method combining remote sensing technology according to claim 1, characterized in that, The step of weighting and fusing the dynamically changing feature set through the feature fusion module to generate a comprehensive risk feature map includes: Assign dynamic weight coefficients to each dynamic feature in the set of dynamic features; Based on the dynamic weighting coefficients, each dynamically changing feature is weighted and concatenated to obtain a preliminary fused feature map; Spatial context enhancement processing is performed on the preliminary fused feature map, and residual connection is performed between the spatial context enhanced feature map and the preliminary fused feature map to generate a comprehensive risk feature map.

4. The geological hazard identification method combining remote sensing technology according to claim 1, characterized in that, The step of generating a disaster early warning signal based on the set of abnormal areas and sending the disaster early warning signal to the disaster monitoring terminal includes: Each abnormal region in the set of abnormal regions is classified into risk levels to generate high-risk, medium-risk, and low-risk region identifiers. Based on the risk level classification results, a corresponding early warning level signal is generated; The warning level signal is associated with the geographic coordinates of the abnormal area to generate a disaster warning signal; The disaster early warning signal is transmitted to the disaster monitoring terminal in real time via a wireless communication network, triggering an early warning prompt from the terminal. The step of classifying each abnormal region in the set of abnormal regions into risk levels to generate high-risk, medium-risk, and low-risk region identifiers includes: Obtain data on the probability of geological disaster risk, the frequency of historical disasters, and the surrounding population density for each abnormal area; The geological disaster risk probability, historical disaster frequency, and surrounding population density data are input into the risk level assessment model, and the risk score for each abnormal area is output. The risk score is divided into high-risk, medium-risk, and low-risk ranges based on a preset risk score threshold. Assign a corresponding risk level label to each abnormal area.

5. The geological hazard identification method combining remote sensing technology according to claim 1, characterized in that, The method also includes a training process for a disaster identification model, including: Collect remote sensing image datasets of historical geological disaster cases and corresponding disaster annotation data to construct a training sample set; The remote sensing image dataset in the training sample set is subjected to terrain feature extraction and dynamic change analysis to generate a set of sample dynamic change features. The dynamic change feature set of the samples is input into the initial geological hazard identification model, and the predicted hazard area is output. Calculate the difference loss between the predicted disaster area and the disaster labeling data, and update the model parameters based on the difference loss; Repeat the training until the model converges to obtain a pre-trained geological hazard identification model.

6. A geological hazard identification system combining remote sensing technology, characterized in that, The method includes a processor and a computer-readable storage medium storing machine-executable instructions, which, when executed by a computer, implement the geological hazard identification method combining remote sensing technology as described in any one of claims 1-5.

Images