Abstract
In the context of global warming, the permafrost regions of the Qinghai-Tibet Plateau have experienced an accelerated warming trend, leading to the emergence of numerous thermokarst geological hazards. These hazards pose significant risks to infrastructure, pastoral activities, and the ecological environment. This study focuses on the period from 2015 to 2021, with the permafrost area of the Qilian Mountains as the primary research site. The study integrates multi-source remote sensing data, including C-band SAR image data from Sentinel-1A, Gaofen-2 (GF-2) satellite images, 30-meter resolution ASTER GDEM digital elevation models, and Landsat images. By combining the Interferometric Synthetic Aperture Radar (InSAR) technology with object-oriented image segmentation and convolutional neural networks for thermokarst hazard detection, and through verification via optical remote sensing interpretation and field surveys, the study successfully identified 51 instances of thermokarst thaw slumps and mudflows. The data obtained provide crucial support for permafrost disaster prevention and mitigation efforts in Qinghai Province. Additionally, the early identification system based on InSAR developed in this study offers valuable insights for monitoring similar thermokarst hazards in the Qilian Mountains and other permafrost regions. This methodological innovation holds considerable significance for advancing research and prevention strategies for thermokarst hazards.