Post classification change detection based on feature-based ensemble classifiers

Abstract

Listen

Translate article

Change detection is a challenging task in the field of remote sensing. Mainly, the change map is used for disaster assessment, monitoring deforestation and urban studies. In this paper, we present a novel method for post classification change detection. Google Earth images of 2011 and 2016 of Bangalore East are used for the study. Multiple features such as texture features, morphological features are extracted using grey level co-occurrence matrix (GLCM) and morphological operations respectively. Linear discriminant analysis (LDA) is used to reduce the dimension of the selected features for the training set. The proposed ensemble classifier system (ECS) exploits K-nearest neighbour (KNN), support vector machine (SVM) and maximum likelihood classifier (MLC). The proposed method adopts the subsample kernel-based subtraction technique to find the difference image; this method greatly reduces data complexity compared to per-pixel-based image subtraction method.