Snow Cover Estimation from Resourcesat-1 AWiFS 每 Image Processing with an Automated Approach

作者：S SUBRAMANIAM; A V SURESH BABU; E SIVASANKAR; V VENKATESHWAR RAO; G BEHERA

来源：International Journal of Image Processing, 2011.

• Snow cover
• Automated algorithm
• Remote Sensing.

Snow and glaciers cover large areas of Himalayas. The resulting runoff from snow and glaciermelt provides nearly 30-50% of the total annual water outlay of most of the rivers in north India.There is a need for continuous monitoring of the Himalayan snow cover area at more frequencyand over longer period in the context of climate change studies. Normalized Difference Snowindex (NDSI) technique which generally used for automated detection of snow cover fromremotely sensed data has limitations in the detection of snow under shadow and exclusion ofwater and gray cloud. A new automated snow cover estimation algorithm has been developedand presented here to overcome the these limitations using the spectral information available inthe Red, Near Infrared and SWIR spectral bands of IRS P6 Resourcesat-1 AWiFS sensor. Theautomated algorithm has been implemented in hierarchical logical steps. Algorithm has beenvalidated by comparing the results obtained with Hall＊s and Kulkarni＊s methods and observed that the new algorithm performs better than other methods in the elimination of noise like glacial waterbodies and water bearing dark clouds, while detecting the snow covered pixels in deep mountainshadows. Satisfactory results have been obtained, when used with several temporal images oflarge image mosaics covering different regions of Himalayas from Kashmir in the west toArunachal Pradesh in the East covering entire snow belt of Indian region. This has shown therobustness of the algorithm in various locations. Intra annual and inter annual snow cover overHimalayas regions of the India was evaluated and the results were encouraging and are suitablefor porting on to a public domain. This algorithm has been evaluated with Landsat ETM and IRSLISS III which has similar spectral bands with different spatial and radiometric resolutions and thealgorithm has been found to be working satisfactorily. The algorithm has been found to be usefulfor regular periodic monitoring of snow cover area and is generic in nature that can be used withvarious sensor data that has similar spectral bands.

• 出版日期2011