Accuracy Analysis of Satellite Image Classification Techniques: Land Cover Changes using Combined LANDSAT and ENVISAT Images

Recent monsoon failures and reduced rain falls urge the environmental and ecology researchers to concentrate on the land cover changes. Significant and efficient way to monitor the land cover changes is satellite image classification. Classification of land cover changes of the study area are identified as used land, unused land, forest and vegetation. Using different kinds of remote sensing data like LANDSAT and ENVISAT, is an important research area for improving the classification performance. This work describes the combination of remotely sensed data, LANDSAT and ENVISAT images, to improve the classification accuracy. Classification algorithms KNN (K-Nearest Neighborhood) and SVM (Support Vector Machine) are tested for the accuracy and KNN in Embedding Space (KNNES) and SVM in Embedding Space (SVMES) are proposed and tested for the improved accuracy. Accuracy is quantified by reporting standard errors i.e., producer accuracy, user accuracy, omission error and commission error.

Sureshkumar N obtained PhD from VIT University. Arun M received PhD from Anna University, Chennai and Post-Doctoral Fellow at University of Aveiro, Portugal. Authors are professors at School of Computing Science and Electronics at VIT University respectively. Their research interests are Satellite Image Processing and Heterogeneous Computing.