Methods And Applications For Glacier Extraction Of Multi-source Remote Sensing Data

Glacier is the largest freshwater resources on the earth and one of the biggest naturalreservoirs of the earth, which has an important influence on the globa l water cycle, regionalwater resources and global climate change. Since the20th century, the global mountainglaciers melting in the generally shrinking state, takes of great scientific interest and research.Glacier change is the response of the climate change and one of the important researchcontents of glaciology, which has an important research value.Remote sensing, as a comprehensive observation technology of global multi-level andmultiple points of view in many areas, forms an obs ervation system from the ground to the air,and space, from the information data collection, processing, analysis and application to theinterpretation, detection and monitoring, which has become an important mean for acquiringthe earth’s resources and environment information. As the increasing number of space-borneremote sensing sensor, the enhancing diversity and quality, remote sensing data become richand diverse, therefore, the application of remote sensing technology in the glacier researchbecome more and more common. Research on extraction of ice based on multi-source remotesens ing image, can not only makes a comprehensive use of the existing remote sens ing imagedata, which serves as a catalyst to meet the increasing and tight investigation and study onglacier area shrinking, but also can play an important role in aspects such as fresh waterresources management and disaster monitoring.This paper focuses on the research of glacier extraction based on multi-source remotesensing data and its application in the field of glacier changing detection, ice mapping and soon. It first reviews the principles and processes of common glacier extraction methods usingoptical remote sensing images. Besides, it proposes a3D (three dimensiona l) visualizationmethod based on least squares surface matching. Furthermore, the ice mapping of SAR(Synthetic Aperture Radar) data has also been researched. Optical remote sensing data andSAR data in Donkemadi district, which locate in the middle of the Thang-la Mountain Rangeof Tibetan Plateau of China, have been used for the test and analysis. The results prove thefeasibility and effectiveness of the proposed methods. The main contents are summarized asfollows:(1)It reviews the principle and process of common glacier extraction methods, includingextraction algorithms based on ratio method, NDSI (normalized difference snow index)method, image segmentation method, classification method and so on. The proposed methodbased on decision-making tree has been put forward to improve the accuracy of extraction. Optical remote sensing data in Dongkemadi has been used for the test and analysis of thismethod.(2) It proposes a3-D visualization method based on least squares surface matching toassess the accuracy of glacier extraction result, which matches the DEM data and opticalremote sensing data in geocoding coordinate. Optical remote sensing data and DEM data inDongkemadi has been used for the test and analysis of this method.(3) Glacier mapping by using SAR data has been carried on after the pre-processingtopographic correction steps including geometric correction and radiometric ratification. Theclassification method has been adopted to serve as catalyst to glacier mapping. To test theproposed method in this paper, a Radarsat-2PolSAR image in C band have been correctedand then applied to map the glacier.