中文摘要：

冰川边界提取是冰川变化监测的一个重要参数，光学冰川遥感方法可以有效、精确的提取出无云区域冰川的边界信息，但是，对于有云区域和天气状况不好的冰川区域却很难获取冰川的边界信息．提出了一种基于SAR干涉失相干特性的提取冰川边界信息的方法，该方法利用内陆冰川表面失相干问题严重的特点，通过比对冰面相干系数和非冰覆盖区域的相干系数，采用阈值法对冰川区域的干涉相干图进行冰川边界的提取．根据上述方法提取冰川边界的精度达到89％，结果表明利用雷达干涉失相干特性进行冰川边界提取是可行的．

英文摘要：

Variation of glacier border is an important parameter of the glacier change monitoring. Although now the optical remote sensing method can extract the variations of glaciers effectively and accurately in the cloudless regions, the variation of glacier with more cloud and bad weather is difficult to be extracted. In this paper, a new method is presented, based on the decorrelation of repeated SAR interferometry, to extract the variation of glacier. This method uses the decorrelation of the inland glacier＇s surface to extract the variation of glacier by comparing the coherence of the glacier and land cover in threshold values. For validation of this method, the classification result is compared with that derived from TM image. An accuracy of better than 89% can be achieved if the classification result from TM image is considered as the ground truth. The results show that this method provides an effective way to identify the glacierized area from the coherent image.