Detection of Environmental Changes through Supervised Classification

Abstract

Nowadays, satellite based remote sensing technology has been successfully utilized for mapping, monitoring and detection of environmental changes. The interested information will be extracted from satellite images by using Digital Image Processing. In this paper, the environmental changes due to natural hazards can be detected and monitored using supervised classification. To extract the information from the multi-date images, the Minimum Distance classifier (MD) is used to identify the classes of images based on RGB color values. It is used in training and also in recognition. The Minimum distance classifier which is based on training data characterizes each class by its mean position on each band. The classification is performed by placing a pixel in the class of the nearest mean. The main purpose of this paper is to evaluate and compare the satellite images of before and after natural disasters in the world with results from the supervised training based methods.