Survey on Machine Learning Approaches for Image Stitching for Panoramic View and Monitoring Global Land Usage using Satellite Images

Image stitching (Mosaicing) is considered as a dynamic research area in computer graphics and computer vision. Image stitching is concerned with concatenating two or more images of the same scene into one high resolution image which is called panoramic image. Image stitching techniques can be categorized into two general approaches: direct and feature based techniques. Direct techniques compare all the pixel intensities of the images with each other, whereas feature based techniques aim to determine a relationship between the images through distinct features extracted from the processed images. The last approach has the advantage of being more robust against scene movement, faster, and has the ability to automatically discover the overlapping relationships among an unordered set of images. The purpose of this paper is to present a survey about the feature based image stitching. The main components of image stitching will be described. A framework of a complete image stitching system based on feature based approaches will be introduced.