Academic Journal of Science and Technology ISSN: 2771-3032 | Vol. 13, No. 2, 2024 276 Potential Disaster Inspection and Analysis Around Reservoirs Based on InSAR Technology Guangchun Liu Liaoning Institute of Science and Technology, Benxi, Liaoning, 117004, China Abstract: Before the occurrence of geological disasters around reservoirs, small deformations occur on the surface or mountains, which are usually difficult to detect. At the same time, geological disasters occur in rainy weather, and traditional monitoring methods and optical satellite monitoring methods are greatly limited. On site monitoring also poses significant safety risks. Therefore, this article uses SAR satellite images and SBAS InSAR method to identify deformation zones in a reservoir and surrounding mountains in Sichuan Province, China, using 58 scenes of Sentinel-1 satellite images from January 2019 to June 2021. The research results indicate that there are 12 areas with significant deformation around the reservoir, with an average deformation rate between -170mm/a and -79mm/a. The cumulative deformation values of P9 and P10 areas exceed 300mm. Combined with geological conditions analysis, these two areas have a certain potential landslide risk and need further monitoring. Keywords: InSAR (Interferometric Synthetic Aperture Radar), SBASSmall Baseline Subset InSAR, Geological hazards, Deformation monitoring. 1. Introduction Sichuan Province in China has numerous mountain ranges and a complex natural environment. It is also located on the Sichuan earthquake zone, with frequent earthquake disasters and intense activity of various fault layers, which are also factors that trigger geological disasters such as landslides, collapses, and mudslides. The formation mechanism of geological disasters is complex and the triggering factors are diverse. Landslide geological disasters often occur with weather conditions such as rainy weather. Optical remote sensing methods are affected by cloud cover and cannot be observed. Traditional monitoring methods require the use of traditional monitoring instruments and equipment for regular observation, with a long observation period. Moreover, deformation analysis can only be conducted on monitoring points, which is not conducive to fully understanding the development and development process of landslide disasters. It is time-consuming and labor-intensive, and on-site monitoring has huge safety hazards. Both personnel and equipment are unsafe [1,2]. As a geological structure with a large water storage capacity, the reservoir in this area poses greater risks, therefore, it belongs to the key monitoring area for geological disasters. [3] . With the rapid development of modern satellite remote sensing technology and Synthetic Aperture Radar (SAR) technology, this technology has the advantages of being fast, efficient, low-cost, high-resolution, wide coverage, high accuracy, and all-weather, which can solve the shortcomings of traditional deformation monitoring. With the continuous development and improvement of technology, modern differential interferometry methods such as DInSAR, Differential InSAR [5,6], PS InSAR, Persistent ScattererInSAR [7,8], SBAS InSAR, Small Baseline Subset InSAR [9,10] have been developed. The method of obtaining long-term small deformations of geological disasters such as landslides has become an effective natural disaster monitoring tool. Yang Chengsheng et al. used multiple SAR data to monitor post earthquake landslide disasters and hidden danger points, and analyzed the distribution and deformation characteristics of landslides. They conducted in-depth analysis of typical landslide bodies and obtained good results [11,12]. Zhao Baoqiang et al. used two time-series InSAR methods, PS and SBAS, to process the 56 Sentinel-1A data and 27 Envisat ASAR data obtained. The research results showed good consistency with the field measurement results in landslide areas [13]. This article uses Sentinel-1 satellite remote sensing data to conduct early identification of landslide disasters in the Reservoir and surrounding mountains of the Heishui River, and provides a detailed analysis of sensitive areas, providing an effective means and method for potential landslide analysis and geological hazard warning in the reservoir area. 2. Methodology 2.1. Principle of Time Series SBAS inSAR The core idea of SBAS InSAR technology is to reorganize and group multi scene Sentinel-1 satellite SAR images based on smaller spatiotemporal baselines, and then perform registration, interferometric generation, and other processing to obtain interferograms with shorter baselines [14]. Due to the short spatiotemporal baseline, the coherence of the generated interferograms is high, which can effectively weaken the impact of decoherence on data accuracy and improve coherence [15]. Connect the isolated interferograms in groups, reduce phase noise by multi view processing of differential interferograms, extract highly coherent pixels, increase the sampling frequency of images, and facilitate the calculation of temporal deformation information. Apply singular value decomposition method and least squares solution in the sense of minimum norm to obtain the surface deformation rate of long time series and the calculation method of deformation time series [16,17]. 2.2. Temporal SBAS InSAR Process and Algorithm 2.2.1. Temporal SBAS InSAR processing flow The temporal SBAS InSAR model is an InSAR processing method that involves impact registration, differential interferometry, filtering, phase unwrapping, and geocoding of 277 SAR satellite images to obtain surface deformation rates, analyze patterns, and reasons. Firstly, select the primary and secondary images (M+1 scenes) with appropriate temporal distribution within the research area, and then use the reference DEM for image registration and interference processing to remove the influence of terrain phase; By estimating the spatiotemporal baseline, it is determined whether the temporal and spatial baselines of the image pair exceed the limit. Then, the differential method is used to generate a small baseline set of interferograms, and each independent interferogram is connected to form an interferogram pair; By using precise POD orbit refinement parameters and ground control points to improve satellite orbit position accuracy, phase unwrapping is performed to remove atmospheric noise and DEM residual noise effects. Finally, geographic coding conversion is performed to convert the radar coordinate system to the ground coordinate system and obtain ground deformation information. The deformation rate and spatiotemporal laws are analyzed, and the causes of deformation are explained and analyzed [18]. The workflow diagram is shown in Figure 1. Figure 1. SBAS InSAR data processing flow 2.2.2. Time series SBAS InSAR algorithm In the processing flow of SBAS InSAR method, statistical analysis is conducted on the coherence coefficient, amplitude deviation and other related parameters of each image pixel, and then the pixels are selected according to the set parameters. The deformation value of the time sequence obtained by phase unwrapping the interference phase [19-20]: T Mdef defdefdefdef t ttt )](, ),(),(),([ 321      (1) Among them: def is accumulate deformation, def ( it ) is the deformation of each image pair. In the SBAS InSAR data processing process, M interferograms are generated, and the matrix composed of the deformation phase of the image at point (x, r) at N time steps is represented by φ (x, r). Therefore, the matrix composed of M interferograms for phase values is: )( idefdef tA   (2) Among them, A is an M×N matrix When M ≥ N, the deformation time series can be obtained through the least squares method. When M