Feature Extraction of Satellite Images Using Decorrelation Stretching and Color Scatter Plots

Abstract

Today we see rapid advancement in the field of Digital image processing and various enhancement techniques have been proposed for the extraction of details in color images. Decorrelation stretching is a method utilized to enhance the color component separation of an image with significant band-band correlation. It is also used as a data visualization tool for improving visual interpretation and make feature discrimination easier. It puts emphasis on high degree of scene variations. The net effect of the process is to obtain an output image whose pixels are well distributed among all possible colors, while preserving the relative sense of hue, saturation, and intensity. By performing linear contrast stretch on the obtained image will enhance the color ranges of the image. For further investigation, the color scatter plot of the image are obtained, which shows how the bands are decorrelated and equalized. The main objective of this paper is to obtain the color separation of images with the help of decorrelation stretching, linear contrast stretch and the color scatter plot of satellite images to investigate the area of vegetation, rock and water content.