Research on Key Technologies and Methods for Early Landslide Detection using Domestic LuTan Satellite InSAR

　　This study focuses on the early identification technology of landslide hazards using domestic LT-SAR satellite images and analyzes typical technical challenges encountered in InSAR-based landslide recognition.
　　First, the importance of multi-look processing of LT-SAR data is discussed, exploring the balance between noise suppression and spatial resolution through multi-look factors. The optimal multi-look factor parameters for LT-SAR images are obtained. Furthermore, the impact of interferogram filtering window on deformation extraction accuracy is analyzed, revealing that the optimal filtering window can effectively suppress interference noise while preserving deformation information. The study also shows that performing atmospheric correction on the InSAR interferogram layer first, followed by terrain-related atmospheric correction, can effectively reduce atmospheric noise and enhance deformation extraction accuracy, avoiding the propagation of phase unwrapping errors. In addition, the paper examines the optimal sequence of InSAR processing steps, finding that correcting external atmospheric errors in the interferogram, followed by orbit error correction, and finally terrain-related atmospheric correction, constitutes the best processing sequence. Finally, a landslide-prone area in the middle and lower reaches of the Minjiang River is used as a case study. Based on optimal and reference parameter sets, early landslide identification experiments using LT-SAR satellite data are conducted, validating the effectiveness and applicability of the optimal parameter set.
　　The results of this study propose an effective InSAR-based landslide recognition strategy for domestic LT-SAR satellite data, contributing to enhancing the application capability of domestic LT-SAR satellite data in early landslide hazard identification and providing valuable references for related scientific research and engineering practice.