Key points are not available for this paper at this time.
Merging photographs from multiple views to create a single panorama is termed Image Stitching. The challenges associated with image stitching encompass issues such as parallax and scene motion, photographs corrupted by noise, managing a vast number of images, and handling high-resolution images. Eliminating noticeable seams is another complex task. Computer-based image stitching is the most prevalent method, involving specialized software that identifies overlapping areas and merges them. Achieving a final image without discernible seams necessitates precise overlaps and consistent exposures. Even if two images are captured in quick succession, differences in lighting can lead to visible seams in the stitched image. The relevance of these seams varies based on the application. In this paper, an automated feature-based approach to image stitching is proposed to create a seamless and expansive panorama. Image feature extraction is done using the Scale-Invariant Feature Transform (SIFT), while feature matching is done using the K-nearest neighbor algorithm. The Random Sample Consensus (RANSAC) algorithm is utilized to calculate homographs for image warping, and a weighted matrix is used for image blending for a smooth panorama creation process. The performance of the proposed method is gauged by comparing the final panorama to an anticipated reference image and calculating the percentage image similarity, which involves the ratio between detected matches and key points.
Gairola et al. (Fri,) studied this question.