Resoluting multispectral image using image fusion and CNN model

Abstract

This paper mainly focus on multispectral image for Entertainment purpose, with such application the quality of the image is an important factor that affects the accuracy of the recognition. Due to hardware limitation, multispectral imaging device may fails to generate high resolution (HR) image. In order to overcome the issue, here we proposes multispectral image super-resolution algorithm (MISR), by fusing low-resolution (LR) multispectral images. In this algorithm the computed response function is used to fuse the multiple images into a single high dynamic range radiance image. It deals with the radiance of the images by mapping. The referred algorithm solves the pre-processing and registration. It uses CNN model to fuse the multiple images into a single high dynamic range image. This fusing technique establishes the common points in the image. Experimental results validate that the MISR algorithm outperforms the state-of-the-arts in terms of both reconstruction accuracy and computational efficiency.