FEATURE BASED AUTOMATIC IMAGE STITCHING

Abstract

Image stitching, also known as mosaicing, is considered a current area of research in computer vision. The goal of image stitching is to create a high-resolution panoramic image by combining two or more photos of the same scene. Feature-based techniques try to create a relationship between the photographs using distinctive features derived from pictures, Feature-based techniques can automatically identify associations between an unordered set of overlapping photos. Within the subject of computer vision, image fusion is considered a current study topic. It has many different algorithms for feature detection and description. This project provided a technique for building a seamless image panorama that uses feature extraction methods to extract visual features. This project compares various feature detection techniques, such as SIFT, ORB, and BRISK. Using feature extractor algorithms, numerous features from both images are extracted. Next, compare the features in both images and stay with the more compatible pairs. For functional matching — KNN (K-Nearest Neighbors) Matcher and BF (Brute Force) Matcher are used and these matchers are examined. The homography was computed using RANSAC algorithm, which stands for Random Sampling Consensus.