青藏高原东南部的海洋性冰川退缩幅度较大。由于受云雨天气的影响，获取高质量的无云光学影像比较难，同时由于地形条件的限制，冰川考察资料也相对较少，迄今为止这些冰川的变化情况仍然是未知数。该项目基于美国解密航片、ERS-2和Sentinel等SAR遥感数据，以及1: 5万数字地形图、SRTM、冰川编目等辅助数据，探索解密航片和SAR数据的冰川识别方法。尤其是探索用Pauli分解和 目标分解的方法如何从多极化SAR影像上将冰川从周围冰碛和其他地物中识别出来。探索基于解密航片立体像对数据，用摄影测量重建冰川表面数字高程模型的方法，以及使用升降轨、多基线的InSAR技术生成冰川表面数字高程模型的方法。用GPS测量和无人机观测数据对遥感获取的冰川面积和数字高程模型进行验证和精度评估。获取长时间跨度的冰川面积和高程变化数据，进而分析冰川储量变化。最后用云的遥感数据产品和再分析数据，探讨冰川变化的原因。

The maritime glaciers retreat dramatically in the southeastern Tibetan Plateau. It is difficult to obtain optical remote sensing images with high quality, which is restricted by cloudy and rainy weather. At the same time, limited by terrain conditions, the field investigations of glaciers are relatively scare, so far the changes of these glaciers are still unknown. On the basis of multi-source remote sensing data, such as USA Declassified Intelligence Satellite Photographs, ESR-2 SAR, Sentinel SAR, as well as ancillary data, including digital topographic maps with a scale of 1:50,000, SRTM and Glacier Inventory data, glacier identification methods are explored based on declassified photographs and SAR data. In particular, a method of Pauli decomposition and target decomposition is explored to identify the glaciers from the surrounding moraine and other objects on the multi-polarization SAR images. Based on stereo pair images of declassified photographs, the method for reconstructing digital elevation model of glacier surface by means of photogrammetry technology is examined. The method for generating digital elevation model of glacier surface by integrating ascending-descending orbit and multi-baseline InSAR technology is also explored. The GPS survey and Unmanned Aerial Vehicle observations is used for validating and evaluating the glacier area and digital elevation model obtained from remote sensing data. Then, the long-term changes of glacier area and elevation are obtained for analyzing the change of glacier volume. Finally, the causes of glacier change are discussed using the remote sensing products of cloud and reanalysis data.