A method for monitoring and predicting surface deformation based on LiCSBAS-InSAR technology

By combining LiCSBAS-InSAR technology and GACOS atmospheric correction technology, an improved CNN model is built, and the genetic algorithm and particle swarm optimization algorithm are used to analyze ground settlement factors and predict future trends, solving the complexity of ground settlement monitoring and prediction in the existing technology. and insufficient prediction problem, high-precision and real-time ground settlement monitoring and prediction are achieved.

CN118587630BActive Publication Date: 2024-11-12KUNMING UNIV OF SCI & TECH
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202410755892.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-12
Publication Date
2024-11-12
Estimated Expiration
2044-06-12

AI Technical Summary

Technical Problem

The existing ground settlement monitoring methods have problems such as data processing complexity, difficulty in correction of atmospheric interference and insufficient prediction. Especially in urban areas, traditional methods are difficult to provide high-precision and real-time ground settlement monitoring and prediction.

Method used

LiCSBAS-InSAR technology combined with GACOS atmospheric correction technology, Sentinel-1 satellite radar image data and multi-source data are used to build an improved CNN model, combine genetic algorithms and particle swarm optimization algorithms to analyze ground settlement factors and predict future trends.

Benefits of technology

It significantly improves the accuracy and prediction accuracy of ground settlement monitoring, provides more continuous and comprehensive monitoring data, promptly warns of potential risks, and provides scientific decision-making support for urban planning and disaster prevention and control.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118587630B_ABST
    Figure CN118587630B_ABST
Patent Text Reader

Abstract

The present invention relates to a method for monitoring and predicting ground subsidence in combination with LiCSBAS-InSAR technology and neural network algorithm, and discloses a method for monitoring and predicting ground deformation based on LiCSBAS-InSAR technology. By utilizing LiCSBAS-InSAR technology and Sentinel-1 satellite data, the ground deformation data of the study area is obtained, and the GACOS atmospheric correction technology is used to optimize data processing, thereby improving the accuracy and reliability of monitoring data. At the same time, in order to predict future ground subsidence trends, this method introduces a CNN model improved based on a genetic algorithm (GA) and a particle swarm optimization (PSO) algorithm, and introduces factors affecting ground subsidence, such as slope and rainfall, into the prediction model, trains the CNN model through a model data set, adjusts the model hyperparameters, and obtains the best model. This method not only improves the efficiency and accuracy of ground subsidence monitoring, but also provides scientific data support and decision-making basis for disaster prevention.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to remote sensing monitoring and data analysis technology, and has broad application prospects in the field of ground subsidence monitoring and prediction, and specifically is a surface deformation monitoring and prediction method based on LiCSBAS-InSAR technology. Background Art

[0002] Land subsidence is a global problem, especially in urban areas, where it poses a direct threat to urban safety due to high population density and complex infrastructure. Land subsidence can be caused by a variety of factors, including groundwater extraction, natural geological activities, building loads, and other human activities. Traditional land subsidence monitoring methods such as leveling, global positioning system (GPS), and optical remote sensing technology have certain limitations, such as being affected by climatic conditions, high monitoring costs, and long data processing cycles.

[0003] In recent years, synthetic aperture radar interferometry (InSAR) technology has attracted widespread attention due to its ability to provide high-precision, high-temporal and high-resolution surface deformation monitoring data. InSAR technology can use the ability of radar signals to penetrate clouds to continuously monitor the surface under any weather conditions. However, although existing time-series InSAR technologies (such as SBAS-InSAR and PS-InSAR) have shown excellent performance in practice, they still face many challenges in data processing and interpretation, such as correction of atmospheric interference, complexity of data integration processing, and high-performance computing requirements for large-scale data processing.

[0004] In addition, although time-series InSAR technology can provide useful data on ground subsidence, InSAR technology alone is still insufficient for predicting future ground subsidence trends. Ground subsidence is affected by a variety of factors, such as geological conditions, climate change, and human activities, and the combined effect of these factors determines the complexity of subsidence. Therefore, a system that can integrate multi-source data and use advanced algorithms to predict future trends is needed. In this case, combining advanced algorithms such as neural networks, especially those models improved by genetic algorithms (GA) and particle swarm optimization (PSO) algorithms, can effectively analyze and process these complex data, not only improving the accuracy of predictions, but also optimizing the monitoring and prediction processes. This system that combines LiCSBAS-InSAR technology and neural network algorithms can effectively address the limitations of existing technologies and provide more reliable decision support for urban planning and disaster prevention. Summary of the invention

[0005] The purpose of the present invention is to provide a method for monitoring and predicting land subsidence. The method comprehensively utilizes LiCSBAS-InSAR technology for time series land subsidence monitoring, and combines with an improved CNN model to analyze and predict the trend of subsidence data, so as to provide more accurate and real-time land subsidence monitoring services, which can reduce the accident rate and property loss rate.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for monitoring and predicting surface deformation based on LiCSBAS-InSAR technology, comprising the following steps:

[0007] Step 1, data collection: using radar image data from the Sentinel-1 satellite as the main input source; determining the main factors affecting land subsidence, and collecting and processing relevant data, including the slope and rainfall in the area;

[0008] Step 2: Use GACOS atmospheric correction technology to optimize the data during data processing to eliminate potential errors caused by the atmosphere and enhance the accuracy and reliability of the data.

[0009] Step 3: Combined with LiCSBAS-InSAR technology and GACOS atmospheric correction products, ground subsidence information can be extracted from preprocessed data. The ground deformation information is processed with grid rainfall data and slope data to obtain a dataset of the main factors affecting ground subsidence.

[0010] Step 4, build a CNN model, introduce the rainfall and slope factor data obtained in step 3 into the CNN neural network improved by combining genetic algorithm (GA) and particle swarm optimization (PSO) algorithm, which is used to analyze and process the factors affecting land subsidence and predict the future subsidence trend.

[0011] Step 5: Train the CNN model using the data set, adjust the model's hyperparameters, and obtain the best model.

[0012] Step 6, model performance evaluation: The performance of the model is comprehensively evaluated through qualitative and quantitative analysis. Based on the CNN model improved by GA and PSO algorithms, the rainfall and slope factors are introduced into the improved CNN model, the predicted results are compared with the actual results, and the prediction accuracy is analyzed and evaluated, and the ground deformation rate in the future is predicted. Based on the analysis results of the neural network model, the development trend of future ground subsidence and potential risk areas are predicted. This method provides a scientific basis for decision makers and helps to formulate more effective preventive measures.

[0013] Preferably, in step 1, radar image data and main ground subsidence factor data of Sentinel-1 satellite are collected, specifically:

[0014] Access the required Sentinel-1 radar imagery data through the European Space Agency (ESA) Copernicus Open Access Hub. Register and use the provided API for automated downloads. Select the Interferometric Wide Swath (IW) mode; rainfall data obtains historical and real-time rainfall data from the Global Precipitation Measurement (GPM) project; slope data uses the DEM data of the area and then processes it using ArcGIS software to obtain slope data.

[0015] Preferably, in step 2, the data is optimized using GACOS atmospheric correction technology, specifically:

[0016] According to the time range of Sentinel-1 radar image data, the corresponding atmospheric delay map (ZTD) was downloaded from the GACOS official website and imported into the SARscape software. The original InSAR phase data was atmospherically corrected using the ZTD map provided by GACOS to eliminate the phase delay caused by atmospheric water vapor changes. The InSAR data before and after atmospheric correction were compared to evaluate the degree of reduction of atmospheric influence.

[0017] Preferably, in step 3, the LiCSBAS-InSAR technology is used to monitor land subsidence and produce a data set of factors that mainly affect land subsidence, specifically:

[0018] InSAR data processing mainly uses LiCSBAS-InSAR technology. Among them, LiCSBAS-InSAR can use the COMETLiCS portal website to download the pre-processed interferogram, and then use the open source LiCSAR data processing software to process the data to obtain ground deformation monitoring data.

[0019] The main factors affecting ground subsidence were determined to be rainfall and slope, and relevant data were obtained and datasets were prepared. Multiple data sources, such as satellite data, radar data, and ground meteorological station data, were combined to obtain ground rainfall data; using digital elevation models, ArcGIS software was used to process DEM data to obtain slope data in the area. The raster rainfall data, slope data, and ground deformation monitoring data were matched with the relevant longitude and latitude using the ArcGIS software multi-value extraction to point tool.

[0020] Preferably, in step 4, a CNN model is built, specifically:

[0021] The CNN model optimized by the GA algorithm is as follows: the hyperparameters of the CNN model (such as the number of convolutional layers, the number of convolutional kernels in each layer, the learning rate and the batch size) are encoded into chromosomes. Then a set of initial populations is randomly generated, and each individual represents a hyperparameter combination. MSE is defined as the fitness function to verify the performance of the model on the test set. The roulette method is used to select the better individuals for mating, and mutation and crossover operations are performed to generate a new generation of populations. The above steps are repeated until the predetermined number of iterations is reached or the stopping condition is met. The optimal hyperparameter combination finally found is used to train the final CNN model and evaluated on the test set.

[0022] The CNN model optimized by the PSO algorithm is: the hyperparameters of the CNN model are represented as the position vector of the particle. A group of particles are randomly generated, their positions and velocities are initialized, MSE is defined as the fitness function, and the performance of the model on the test set is verified. Then the best position found by each particle (individual extreme value) and the best position of the entire particle group (global extreme value) are recorded. Then the speed and position are updated using the following formula

[0023]

[0024] Where i = 1, 2, ..., N, N is the total number of particles in this group; is the velocity of particle i at the current iteration t, v i The maximum value is V max >0, if V max =V i ; rand() is a random number between (0, 1); is the position of particle i at the current iteration t; c 1 and c 2 is the learning factor, 1 =c 2 =2.

[0025] The new fitness value of each particle is calculated, and the individual extremum and global extremum are updated. The speed and position updates are repeated until the predetermined number of iterations is reached or the stopping condition is met.

[0026] Preferably, in step 5, the CNN model is trained and the hyperparameters of the model are adjusted to obtain the best model, specifically:

[0027] The datasets matching rainfall, slope and ground deformation data with relevant longitude and latitude were divided into training set, validation set and test set by using ArcGIS software multi-value extraction to point tool, and CNN model was constructed using ResNet network structure, Sigmoid was selected as activation function and MSE was selected as loss function. The dataset was input into CNN model, and the learning rate, batch size, convolution kernel size, number of convolution layers and weight decay hyperparameter data were adjusted, and the weights of the network were updated by gradient descent to obtain the maximum hyperparameter combination to obtain better performance and generalization ability.

[0028] Preferably, in step 6, the model performance evaluation is specifically:

[0029] The root mean square error (RMSE), mean absolute error (MAE), mean square error (MSE) and coefficient of determination (R 2 ) measurement indicators to evaluate the accuracy of the model. Among them, RMSE is a standard measure of the difference between the predicted value and the actual value, and the calculation formula is:

[0030]

[0031] Among them, y i is the actual value, is the predicted value, and n is the number of samples. The smaller the RMSE value, the higher the accuracy of the model.

[0032] MAE measures the average absolute value of the difference between the predicted value and the actual value. The smaller the value, the higher the prediction accuracy of the model. The formula is:

[0033]

[0034] MSE is the average of the squares of the differences between the predicted values ​​and the actual values, and its calculation formula is:

[0035]

[0036] R 2 It measures the amount of variation explained by the model and reflects the degree of fit of the model to the data. The calculation formula is:

[0037]

[0038] in, is the average of the actual values. 2 The closer the value is to 1, the stronger the explanatory power of the model is and the higher the prediction accuracy is.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] Through the above technical solution, the present invention can significantly improve the monitoring efficiency and prediction accuracy of ground subsidence, especially in complex environments. Compared with traditional ground subsidence monitoring methods, this system can provide more continuous and comprehensive monitoring data. At the same time, by using advanced data processing and analysis technology, it can timely predict and warn possible ground subsidence risks, and provide scientific decision-making for disaster prevention and control. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a specific flow chart of the present invention;

[0042] Figure 2 The rainfall data and DEM data diagram of the embodiment of the present invention;

[0043] Figure 3 This is a STD variation diagram of the GACOS atmospheric correction product before and after correction in an embodiment of the present invention;

[0044] Figure 4 Flowchart for LiCSBAS-InSAR data processing;

[0045] Figure 5 This is a LiCSBAS-InSAR result diagram of an embodiment of the present invention;

[0046] Figure 6 This is a diagram of the training process of the InSAR-improved CNN model according to an embodiment of the present invention;

[0047] Figure 7 This is a comparison chart of the time series results before and after the prediction of the embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] See also Figure 1-Figure 7 A method for monitoring and predicting ground deformation based on LiCSBAS-InSAR technology comprises the following steps:

[0050] Step 1, data collection: using radar image data from the Sentinel-1 satellite as the main input source; determining the main factors affecting land subsidence, and collecting and processing relevant data, including the slope and rainfall in the area;

[0051] Step 2: Use GACOS atmospheric correction technology to optimize the data during data processing to eliminate potential errors caused by the atmosphere and enhance the accuracy and reliability of the data.

[0052] Step 3: Combined with LiCSBAS-InSAR technology and GACOS atmospheric correction products, ground subsidence information can be extracted from preprocessed data. The ground deformation information is processed with grid rainfall data and slope data to obtain a dataset of the main factors affecting ground subsidence.

[0053] Step 4, build a CNN model, introduce the rainfall and slope factor data obtained in step 2 into the CNN neural network improved by combining genetic algorithm (GA) and particle swarm optimization (PSO) algorithm, which is used to analyze and process the factors affecting land subsidence and predict the future subsidence trend.

[0054] Step 5: Train the CNN model using the data set, adjust the model's hyperparameters, and obtain the best model.

[0055] Step 6, model performance evaluation: The performance of the model is comprehensively evaluated through qualitative and quantitative analysis. Based on the CNN model improved by GA and PSO algorithms, the rainfall and slope factors are introduced into the improved CNN model, the predicted results are compared with the actual results, and the prediction accuracy is analyzed and evaluated, and the ground deformation rate in the future is predicted. Based on the analysis results of the neural network model, the development trend of future ground subsidence and potential risk areas are predicted. This method provides a scientific basis for decision makers and helps to formulate more effective preventive measures.

[0056] In this embodiment: in step 1, radar image data and main ground subsidence factor data of Sentinel-1 satellite are collected, specifically:

[0057] Access the required Sentinel-1 radar imagery data through the European Space Agency (ESA) Copernicus Open Access Hub. Register and use the provided API for automated downloads. Select the Interferometric Wide Swath (IW) mode; rainfall data obtains historical and real-time rainfall data from the Global Precipitation Measurement (GPM) project; slope data uses the DEM data of the area and then processes it using ArcGIS software to obtain slope data.

[0058] In this embodiment: In step 2, the GACOS atmospheric correction technology is used to optimize the data, specifically:

[0059] According to the time range of Sentinel-1 radar image data, the corresponding atmospheric delay map (ZTD) was downloaded from the GACOS official website and imported into the SARscape software. The original InSAR phase data was atmospherically corrected using the ZTD map provided by GACOS to eliminate the phase delay caused by atmospheric water vapor changes. The InSAR data before and after atmospheric correction were compared to evaluate the degree of reduction of atmospheric influence.

[0060] In this embodiment: In step 3, the LiCSBAS-InSAR technology is used to monitor land subsidence and produce a data set of factors that mainly affect land subsidence, specifically:

[0061] InSAR data processing mainly uses LiCSBAS-InSAR technology. Among them, LiCSBAS-InSAR can use the COMETLiCS portal website to download the pre-processed interferogram, and then use the open source LiCSAR data processing software to process the data to obtain ground deformation monitoring data.

[0062] The main factors affecting ground subsidence were determined to be rainfall and slope, and relevant data were obtained and datasets were prepared. Multiple data sources, such as satellite data, radar data, and ground meteorological station data, were combined to obtain ground rainfall data; using digital elevation models, ArcGIS software was used to process DEM data to obtain slope data in the area. The raster rainfall data, slope data, and ground deformation monitoring data were matched with the relevant longitude and latitude using the ArcGIS software multi-value extraction to point tool.

[0063] In this embodiment: in step 4, a CNN model is built, specifically:

[0064] The CNN model optimized by the GA algorithm is as follows: the hyperparameters of the CNN model (such as the number of convolutional layers, the number of convolutional kernels in each layer, the learning rate and the batch size) are encoded into chromosomes. Then a set of initial populations is randomly generated, and each individual represents a hyperparameter combination. MSE is defined as the fitness function to verify the performance of the model on the test set. The roulette method is used to select the better individuals for mating, and mutation and crossover operations are performed to generate a new generation of populations. The above steps are repeated until the predetermined number of iterations is reached or the stopping condition is met. The optimal hyperparameter combination finally found is used to train the final CNN model and evaluated on the test set.

[0065] The CNN model optimized by the PSO algorithm is: the hyperparameters of the CNN model are represented as the position vector of the particle. A group of particles are randomly generated, their positions and velocities are initialized, MSE is defined as the fitness function, and the performance of the model on the test set is verified. Then the best position found by each particle (individual extreme value) and the best position of the entire particle group (global extreme value) are recorded. Then the speed and position are updated using the following formula

[0066]

[0067] Where i = 1, 2, ..., N, N is the total number of particles in this group; is the velocity of particle i at the current iteration t, v i The maximum value is V max >0, if V max =V i ; rand() is a random number between (0, 1); is the position of particle i at the current iteration t; c 1 and c 2 is the learning factor, 1 =c 2 =2.

[0068] The new fitness value of each particle is calculated, and the individual extremum and global extremum are updated. The speed and position updates are repeated until the predetermined number of iterations is reached or the stopping condition is met.

[0069] In this embodiment: In step 5, the CNN model is trained, and the hyperparameters of the model are adjusted to obtain the best model, specifically:

[0070] The datasets matching rainfall, slope and ground deformation data with relevant longitude and latitude were divided into training set, validation set and test set by using ArcGIS software multi-value extraction to point tool, and CNN model was constructed using ResNet network structure, Sigmoid was selected as activation function and MSE was selected as loss function. The dataset was input into CNN model, and the learning rate, batch size, convolution kernel size, number of convolution layers and weight decay hyperparameter data were adjusted, and the weights of the network were updated by gradient descent to obtain the maximum hyperparameter combination to obtain better performance and generalization ability.

[0071] In this embodiment: in step 6, the model performance evaluation is specifically:

[0072] The root mean square error (RMSE), mean absolute error (MAE), mean square error (MSE) and coefficient of determination (R 2) measurement indicators to evaluate the accuracy of the model. Among them, RMSE is a standard measure of the difference between the predicted value and the actual value, and the calculation formula is:

[0073]

[0074] Among them, y i is the actual value, is the predicted value, and n is the number of samples. The smaller the RMSE value, the higher the accuracy of the model.

[0075] MAE measures the average absolute value of the difference between the predicted value and the actual value. The smaller the value, the higher the prediction accuracy of the model. The formula is:

[0076]

[0077] MSE is the average of the squares of the differences between the predicted values ​​and the actual values, and its calculation formula is:

[0078]

[0079] R 2 It measures the amount of variation explained by the model and reflects the degree of fit of the model to the data. The calculation formula is:

[0080]

[0081] in, is the average of the actual values. 2 The closer the value is to 1, the stronger the explanatory power of the model is and the higher the prediction accuracy is.

[0082] In summary: The method of the present invention can effectively improve the accuracy and prediction ability of ground subsidence monitoring by integrating advanced remote sensing technology and machine learning algorithms, which helps to identify potential risks early and provide important decision-making support for safety management.

[0083] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0084] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring and predicting surface deformation based on LiCSBAS-InSAR technology, characterized in that: The following steps are involved: Step 1, data collection: using radar image data from the Sentinel-1A satellite as the main input source; determining the main factors affecting land subsidence, and collecting and processing relevant data, including slope and rainfall in the monitoring area; Step 2: Use GACOS atmospheric correction technology to optimize the data in data processing to eliminate potential errors caused by the atmosphere; Step 3: Combining LiCSBAS-InSAR technology and GACOS atmospheric correction products, the ground subsidence information can be extracted from the preprocessed data, and the ground deformation information is processed with the grid rainfall data and slope data to obtain the data set of the main factors affecting the ground subsidence; Step 4: Build a CNN model. The rainfall and slope factor data sets obtained in step 3 are introduced into the CNN neural network improved by combining genetic algorithm, i.e., GA, and particle swarm optimization, i.e., PSO algorithm, which is used to analyze and process the factors affecting land subsidence and predict future subsidence trends. The parameters of the CNN model are optimized by genetic algorithm and particle swarm optimization algorithm. Specifically: The CNN model optimized by the GA algorithm is as follows: the hyperparameters of the CNN model, including the number of convolutional layers, the number of convolutional kernels in each layer, the learning rate and the batch size, are encoded into chromosomes, and then a set of initial populations is randomly generated, each individual represents a hyperparameter combination, and the MSE is defined as the fitness function. The performance of the model on the test set is verified, and the roulette method is used to select the better individuals for mating, and mutation and crossover operations are performed to generate a new generation of populations. The above steps are repeated until the predetermined number of iterations is reached or the stopping condition is met. The optimal hyperparameter combination finally found is used to train the final CNN model and evaluated on the test set. The CNN model optimized by the PSO algorithm is as follows: the hyperparameters of the CNN model are represented as the position vector of the particle, a group of particles are randomly generated, their positions and velocities are initialized, MSE is defined as the fitness function, the performance of the model is verified on the test set, and then the best position found by each particle, i.e., the individual extreme value, and the best position of the entire particle group, i.e., the global extreme value, are recorded, and then the speed and position are updated using the following formula: v i (t+1) =v i (t) +c1×rand()×(pbest i -x i (t) )+c2×rand()×(gbest i -x i (t) ) x i (t+1) =x i (t) +v i (t+1) Where i = 1, 2, ..., N, N is the total number of particles in this group; v i (t) is the velocity of particle i at the current iteration t, v i The maximum value is V max >0, if V max =V i ; rand() is a random number between (0, 1); x i (t) is the position of particle i at the current iteration t; c1 and c2 are learning factors, generally c1=c2=2; Calculate the new fitness value of each particle, update the individual extreme value and the global extreme value, and repeat the speed and position update until the predetermined number of iterations is reached or the stop condition is met; Step 5: Train the CNN model with the data set, adjust the model's hyperparameters, and obtain the best model. Train the CNN model with the data set, adjust the model's hyperparameters, and obtain the best model. Specifically: The ArcGIS software multi-value extraction to point tool is used to divide the dataset that matches the rainfall, slope and ground deformation data with the relevant longitude and latitude into training set, validation set and test set, and the ResNet network structure is used to build the CNN model. Sigmoid is selected as the activation function and the mean square error is used as the loss function. The dataset is input into the CNN model. The learning rate, batch size, convolution kernel size, number of convolution layers and weight decay hyperparameter data are adjusted, and the network weights are updated by gradient descent to obtain the largest hyperparameter combination for better performance and generalization ability.

2. The method for monitoring and predicting surface deformation based on LiCSBAS-InSAR technology according to claim 1, characterized in that: Data collection includes radar image data from the Sentinel-1A satellite, digital elevation model of the study area and its rainfall data.

3. The method for monitoring and predicting surface deformation based on LiCSBAS-InSAR technology according to claim 1, characterized in that: The GACOS atmospheric correction technique described in step 2 optimizes data processing, specifically: Collect atmospheric data including temperature, humidity and pressure information; establish an atmospheric model, including atmospheric delay and atmospheric humidity parameters; finally, perform atmospheric correction and optimize the interferometric image. The use of GACOS atmospheric correction technology can effectively improve the efficiency and accuracy of InSAR data processing and provide reliable data support for ground deformation monitoring and prediction.

4. The method for monitoring and predicting surface deformation based on LiCSBAS-InSAR technology according to claim 1, characterized in that: In step 3, InSAR data processing and production of the main factors affecting ground subsidence datasets are as follows: InSAR data processing mainly uses LiCSBAS-InSAR technology to determine that the main factors affecting ground subsidence are rainfall and slope, and obtain relevant data and make a data set, combining multiple data sources, including satellite data, radar data and ground meteorological station data to obtain ground rainfall data; Using the digital elevation model, ArcGIS software was used to process DEM data to obtain slope data of the area, and the raster rainfall data, slope data, and ground deformation monitoring data were matched with the relevant longitude and latitude using the ArcGIS software’s Extract Multi-Value to Point tool.

5. The method for monitoring and predicting surface deformation based on LiCSBAS-InSAR technology according to claim 1, characterized in that: In step 5, the model performance is evaluated, specifically: The performance of the model was evaluated using the test set, and the model accuracy was evaluated using any measurement indicators such as root mean square error, mean absolute error, mean square error, and coefficient of determination.

Citation Information

Patent Citations

  • Landslide deformation monitoring and early warning method based on SAR data

    CN111474544A

  • Landslide susceptibility evaluation method and system

    CN114091274A

  • InSAR-based debris flow disaster susceptibility dynamic evaluation method

    CN116486279A

  • Multi-factor landslide early warning monitoring method based on InSAR technology

    CN116930960A

  • Apparatus and method for comprehensively monitoring slopes based on wireless network

    US20140159915A1