CNN based change detection for urban imagery

Identification of changes in the green cover of urban settlements and keeping them in check has become obligatory because of the eminent dangers of climate change and pollution. The exact location, accurate identification of topographic features and the extraction of the required parameters for their identification form a basis for change detection in the region of interest. In this project to support object-based classification, the required spectral band features are obtained. The extracted features are then used for computing the vegetation index based on which the images are labelled and are then used to train the CNN classifier. The percentage difference in vegetation index is then measured and the image is classified accordingly.