Comparison of glacier change detection using pixel based and object based classification techniques

Glaciers are important indicators of sustainable life on the globe by various means and hence provide a good motivation for continuous monitoring. Temporal analysis of two valley glaciers (namely Apsara and Singhi lie in Shaksgam valley, China) have been performed using supervised, knowledge and object based classification techniques. Landsat MSS and TM data from 1978 to 2011 have been used. It has been observed that object based and supervised classifications are relatively more effective than knowledge based classification to detect the glaciers change having 95%, 93% and 86% overall accuracy respectively. Variation in glaciers extent is due to their characteristics relating to topography, geographic condition, orientation, altitude as well as local climate conditions. Apsara and Singhi glacier lost 3.32 and 8.98 percent of their area respectively throughout the study period.