Segmentation Algorithm for Image Fusion using Watershed Transform

Abstract

Segmentation algorithm plays a vital role in region based image fusion process. Features should be segmented as single regions. Feature may split into more than one region and each region has to be treated separately. If possible, less number of regions should be generated to reduce the computational burden. Segmentation can be done either separately (uni-model segmentation) or jointly (joint segmentation). Uni-model segmentation method may create many regions than joint segmentation since different images have different features. The challenge of finding small targets in big images lies in the characterization of the background clutter. In this paper we report that it is possible to identify or mark the foreground objects and background locations by using watershed transform. By computing the segmentation function, foreground markers and background markers, we can achieve the modified segmentation function so that it has minima at the foreground and background marker locations.