Accuracy Evaluation Of Spaceborne InSAR Deformation Analysis Based On GNSS Monitoring Dat

In recent years,geological disasters have seriously threatened the safety of people’s lives and property,becoming a major issue faced by public safety governance.How to detect potential hazards in the early stage is the key to geological disaster prevention and control,among which surface deformation monitoring is an effective means of identifying geological hazards.At present,there are many technical means for monitoring surface deformation,with traditional measurement methods mainly represented by leveling and GNSS measurement.However,traditional measurement has long observation cycles,high monitoring costs,low efficiency and spatial resolution,making it difficult to conduct large-scale monitoring.High spatial resolution and wide coverage radar remote sensing observation data can effectively compensate for the shortcomings of the above observations.However,due to factors such as satellite orbit positioning accuracy and atmospheric phase delay error,its time resolution and ground monitoring accuracy are relatively low,and may even lead to erroneous conclusions,leading to doubts about the accuracy of some monitoring results.This article aims to construct a reliable In SAR accuracy evaluation method and qualitatively and quantitatively evaluate the In SAR deformation accuracy results based on GNSS monitoring data.The main research and achievements are as follows:(1)Taking the In SAR results of Kunyang phosphate mine as the accuracy evaluation object,the SBAS-In SAR technology was used to process 17 Sentinel-1A data from January 2021 to December 2021,and the same period GNSS data was used to compare and verify the In SAR monitoring results from two aspects: average rate and sequential displacement.The evaluation results indicate that during the monitoring time period,the root mean square error of the linear average rate between the two is14.3mm/a,and the accuracy of In SAR is further improved to 11.4mm/a through reference reference compensation in the later stage;The average time series root mean square error of 16 GNSS points is 12.5mm,and the In SAR monitoring accuracy can reach the millimeter level.(2)Taking the In SAR results of the Xin Jian landslide at the Ahai Power Station as the accuracy evaluation object,the SBAS-In SAR technology was used to process 56Sentinel-1A data from July 2017 to April 2019,and the same period GNSS data was used to compare and verify the In SAR monitoring results from two aspects: average rate and sequential displacement.The evaluation results indicate that during the monitoring time period,the root mean square error of the linear average rate of the two is 22.9mm/a,and the accuracy is further improved by compensating the In SAR with reference benchmarks,reaching 22.1mm/a;The root mean square error of the average temporal displacement of 16 GNSS monitoring stations is 25.3mm.The In SAR technology and GNSS monitoring results are basically consistent in areas with small shape variables,and the monitoring accuracy is relatively high;However,in areas with severe deformation,there is still a certain deviation from the GNSS monitoring results,and the monitoring accuracy still needs to be improved.(3)Taking the In SAR results in Lincang City as the accuracy evaluation object,42Sentinel-1A images covering the research area from July 1,2018 to May 22,2021 were processed using SBAS-In SAR technology,and spatial interpolation was used to fit the missing In SAR data in Lincang City.Finally,the same period GNSS monitoring data was obtained to evaluate the accuracy of SBAS-In SAR monitoring results.Research has shown that the root mean square errors of the Kriging interpolation method and the inverse distance weighted interpolation method are 8.5mm and 9.7mm,respectively,indicating high consistency between GNSS data and In SAR monitoring values.Finally,the accuracy of In SAR monitoring was further improved through benchmark deviation compensation,reaching 7.0mm and 8.5mm,and the Kriging interpolation method had better performance than inverse distance weighted interpolation.