根据模型和分布函数,本文首先依据多年平均气温、地温和SRTM等数据对研究区域冰缘地貌的分布范围进行分别提取,并利用遥感数据和人工解译方式对其进行了修正。在此基础上,采用一定指标,利用SRTM数据对冰缘地貌次级类型(如起伏度、海拔高度和坡度等)进行了提取,从而完成研究区域冰缘地貌信息的提取。研究结果表明:①研究区域冰缘地貌总面积约5.15×10⁴km²,主要分布在研究区域的西北部和西南部,另外在东北部也有少量分布;通过提取,研究区域中最重要的冰缘地貌类型是冰缘作用的中起伏缓极高山,面积约0.82×10⁴km²,分布范围较广。②冰缘地貌的分布与海拔高度、气温和地温等有密切的关系,基于此提取的结果可为冰缘地貌的解译提供一定的参考;由于青藏高原气象站点较少,数据精度较低,自动提取精度受到很大限制,因此进行人工解译修正是非常重要和必不可少的。

Qinghai-Tibet Plateau located in southwestern China is one of the important geomorphological units of the country's terrestrial part. Because of its high altitude, vast area, and the mid-latitude location, known as "the third pole", it has close correlation with the biggest monsoon system on the globe which has not only sensitive responding character but significant impact on the global climate change. Hence it becomes one of the hot spots of research. Specific geographical environment, unique sea level elevation and frigid climate condition of the Qinghai-Tibet Plateau make kinds of periglacial geomorphology brand into the geomorphological landscape of the Plateau. Taking the district of 1 ∶ 1,000,000 international standard of Lhasa map sheet (serial number H-46) as an example, this paper explores an information extraction method of periglacial geomorphology on the Qinghai-Tibet Plateau based on multiple source data such as remote sensing data, SRTM, air temperature and ground temperature. In the research, the primitive boundary of periglacial geomorphology is acquired by the indexes such as annual mean temperature of national standard station with a resolution of 1km, annual mean ground temperature and elevation through models. The bound is revised and synthetically processed by using the features such as color, shape and texture of remote sensing images (TM and ETM+ of Landsat). Hence the extent of periglacial geomorphology of the study area is determined. Then the morphological indexes such as relief, elevation and slope of the periglacial geomorphology in the study area are computed by using the SRTM-DEM data. Based on the features such as the completeness of geomorphological units, in virtue of geomorphological expert knowledge and the features of remote sensing images, the indexes are revised by using man-computer mutual intelligentized extraction method.At length quantificational exaction of morphological indexes of periglacial geomorphology in the study area is completed. Finally, the morphological characters of periglacial geomorphology in the study area are integrated and the semi-auto matic remote sensing interpretation result map of periglacial geomorphology in the study area and statistical attribute data are achieved. This research can accomplish remote sensing precise location of geomorphological unit boundary and exact attribute evaluation of geomorhpological types based on multiple source data, which promotes the development of extraction methods of remote sensing geomorphological information. Thus the research method can extend to other areas of the Qinghai-Tibet Plateau and is important in theory and practice.