Image fusion using direction wavelet approach and performance evaluation of multiple source images

Abstract

Image fusion is the approach to get one composite output image from number of variety of input images. We can fuse images which are out of focus, here images of different focal length are considered as input whereas output image is expected to be clear and visible. There are various types of sensors available today having wide variety of focal length and hence they generate images in which some objects are clear some are not. Again multi-sensor devices provide huge data to analyze and interpret. So we require some sort of mechanism to fuse all this data into single one. Therefore there is a basic requirement of certain image fusion algorithm which will perform the desired task and gives some accurate and reliable information about the scene. In the research work a method is developed to perform the fusion of multiple images to get a single image which is more informative, reliable and accurate. The proposed method is based on complex wavelet transform with the combination of direction wavelet decomposition method. The method is good at direction selectivity, also perfect reconstruction which is very important aspect for fusion of images. The final fused image is enhanced in such a way that it is useful for interpretation and analysis and assists to come up with certain decision about the scene.