图像拼接
视差
计算机科学
计算机视觉
人工智能
图像(数学)
计算机图形学(图像)
算法
作者
Xinrui Zhang,Shengwei Guo,Guobing Sun
标识
DOI:10.1109/cipcv61763.2024.00012
摘要
Image stitching is the process that can create a broad visual image by combining angular differences images. Stitching results are typically generated in two steps: (i) aligning images employing image registration to provide preliminary stitched images with a wide field and (ii) optimizing the preliminary stitched images by image blending, which can unify brightness and color differences and eliminate the stitching seams. In addition, to enhance stability and generalization, the image stitching algorithm primarily applies the natural image as an input. This paper focuses on the process of stitching for natural images with parallax, summarizes the image stitching algorithm from the critical steps, including image registration and image blending, and discusses its merits and limitations.
科研通智能强力驱动
Strongly Powered by AbleSci AI