Landslide disasters pose a continuous threat to the safety of life and property of mountain residents. Especially under heavy rainfall, landslide areas are often accompanied by severe surface deformation, resulting in serious incoherence of synthetic aperture radar interference data, which restricts the deformation monitoring capability of traditional timing InSAR technology in this type of area. To address this problem, this paper takes the large landslide that occurred in Nanyu township, Zhouqu county, Gansu province in July 2018 as the research object, and prensents the distributed scatterer InSAR (DS-InSAR) deformation monitoring method based on statistical homogeneity detection and time phase link optimization. This method introduces a statistical homogeneity determination model under the traditional DS-InSAR framework, and identifies clusters of pixels with stable scattering characteristics through non-parametric testing, thereby expanding the monitoring coverage of low-coherence areas. At the same time, the shortest path constraint strategy is used to optimize the interferogram network structure to improve connection density and temporal coherence. It is also combined with a time-Phase Linking joint estimation algorithm to optimize the interference phase as a whole, enhancing the temporal continuity and stability of deformation information. The research results show that this method significantly improves the distribution density and temporal coherence of monitoring points in the collapse area of Nanyu landslide. The number of effective coherent points in creased from 23 232 to 43 463, an increase of about 87%, and the average posterior coherence coefficient was improved by 0.3, which effectively revealing the dynamic evolution characteristics of the landslide main body. The optimized annual average deformation rate chart shows that the maximum sliding rate of the main deformation area of the landslide exceeds 70 mm/a, and the timing curve of the characteristic point is highly consistent with the rainfall process. This study verifies the applicability and effectiveness of the timing phase optimization method in high-voltage landslide areas, and provides strong technical support for InSAR monitoring of landslides in complex mountainous areas.

landslide monitoring  /  coherence optimization  /  timing InSAR  /  Phase Linking  /  Nanyu landslide  /  deformation extraction