Assessment of different CVA based change detection techniques using MODIS dataset

Abstract

Change Vector Analysis (CVA) as change detection technique has useful capabilities of extracting and identifying land cover changes in terms of change-magnitude and change-direction from two different temporal satellite imageries. Since past two-three decade, many effective CVA based change detection techniques, e.g., Improved Change Vector Analysis (ICVA), Modified Change Vector Analysis (MCVA) and Change Vector Analysis Posterior-probability Space (CVAPS), have been developed to overcome the difficulty that exists in CVA. But the choice of best suitable CVA technique for particular area is a very difficult process because different CVA techniques have their own features and no single technique is applicable to all case studies. An efficacy of aforementioned CVA techniques has not been examined on snow cover area of rugged terrain. On the other hand, topographic distortions such as shadow, affects the performance of change detection analysis because hilly surface slope towards the sun receiving more reflectance value as compared to slope opposite direction from the sun. It suppresses the vital information that leads to the inaccurate consequences. So topographic corrections are also necessary to be executed on satellite dataset before further considerations. In the present paper, different CVA techniques have been investigated over snow covered area of rugged terrain using topographic corrected MODIS dataset to find out the best possible technique which could distinguish more accurately changed and no-changed pixels, and also accurately perform “from-to” change detection. Based on limited study done in this paper, it is formed that CVAPS technique has greater potential than MCVA and ICVA techniques to evaluate the overall transformed data over snow covered area of rugged terrain. Results of this study are expected to be potentially useful for more accurate analysis of LULC changes over rugged terrain which will, in turn, improve the utilization of MODIS dataset for such applications by various users.