关键词: 尾矿库, 分布式目标, InSAR, 变形监测

Abstract: Tailings pond is a dangerous source of artificial high gravity potential energy. Once the dam breaks,it will seriously pollute the environment and threaten the life and property safety of downstream people. However,the existing deformation monitoring methods of tailings pond have the problems of time-consuming,laborious and high cost. Therefore,this paper propose a deformation information extraction method of tailings pond based on distributed target interferometric radar measurement (DSInSAR),which is helpful to realize high-efficiency and high-precision deformation monitoring of tailings pond. This method uses the FaSHPS (Fast SHP Selection) method based on the parameter hypothesis test to select homogeneous pixels,extracts the dominant scatterer phase through the eigenvalue decomposition of the coherence matrix to achieve phase optimization,and then obtains the time-series deformation information of the tailings pond in combination with the permanent scatterer ( PS). Taking 23 Sentinel-1B images as data sources,this method is used to obtain the time-series surface deformation information of Brumadinho tailings pond in Brazil. The results show that the density of monitoring points of tailings pond obtained by this method is 2. 2 times that of PS-InSAR,and the deformation rate obtained by the two methods has a strong correlation,and the deformation of tailings pond shows a trend of outward diffusion from the middle of dam crest. The correlation analysis between rainfall and deformation shows that rainfall has a significant impact on the deformation of tailings pond. Generally,there are few deformation in dry season,but the opposite in rainy season.