Effective segmentation technique for 0.5m resolution World View-1 imagery

Over the years, scientists have been trying to formulate efficient segmentation techniques to characterize the objects in the image. Recent days segmentation is used extensively in automated image analysis. The ultimate aim in an automated system is to extract important features from the image data, from which description, interpretation or understanding of the scene can be provided by the machine. This task becomes difficult while dealing with high resolution imagery due to superfluous details, too many objects and possibly low contrast ratio present in the imagery. In this paper, we discuss the effectiveness of object based segmentation method such as edge guided region growing over simple region growing segmentation techniques for World View-0.5 meter high resolution imagery (Digital Globe imagery products). This paper is an attempt to develop an unsupervisory algorithm for automated software to segment high resolution urban satellite images. Experimental results demonstrate promising performance achievements compared to simple region based segmentation methods for detecting tall buildings. Furthermore, the simplicity of this method is an attractive feature for real-time applications.