Image processing application: Overlapping of Images for faster video processing devices

Abstract

This paper mainly concentrates on the merging and impulse noise removal technique in images. Image fusion is the operation of merging a pair or more images into an individual image while maintaining the main property of each image. Five dissimilar filtering algorithms are used independently for refining the image recorded by the sensor. The filtered images are merged to get a favourable yet more standard image out of independently denoised images. In-order to get higher quality of image with noise withdrawing performance of the proposed merging technique from human’s perspective, an edge detection is carried out using a canny filter to the merged image. Any value of salt and pepper noise is reduced by using these five filters. Fact-finding outcome display that this process has the ability of putting together effective outcomes of distinct images into single images. This process is algorithmically straight forward and can be used for merging any number of images and in real time applications too. Image Fusion is a top-tier procedure used in remote sensing, satellite vision, military areas, perfect for the human vision, robotics field and medical applications. In the proposed system, firstly, an image is recorded by a sensor. Secondly, the image is filtered in equivalent five different creamy filters. The denoised images are achieved from five mismatched filters are merged them at third level and in the final level, canny filter is applied to the fusion of recovered image to acquire a giant standard image free from impulse noise. MATLAB tool is used for the implementation of the proposed image fusion technique.