An improved adaptive triangular mesh-based image warping method

• DOI: 10.3389/fnbot.2022.1042429
• Publication Date: 2023-01-23T05:47:05Z
• Authors: Wei Tang; Fangxiu Jia; Xiaoming Wang
• Keywords: Image stitching; Panorama; Image warping; Ghosting; Distortion (music)

It is of vital importance to stitch the two images into a panorama in many computer vision applications motion detection and tracking virtual reality, panoramic photography, tours. To preserve more local details with few artifacts panoramas, this article presents an improved mesh-based joint optimization image stitching model. Since uniform vertices are usually used warps, we consider matched feature points as grid strengthen constraints on deformed vertices. Simultaneously, define energy function add color similarity term perform alignment. In addition good alignment minimal distortion, regularization parameter strategy combining our method as-projective-as-possible (APAP) warp introduced. Then, controlling proportion each part by calculating distance between vertex nearest point vertex. This ensures natural effect non-overlapping areas. A comprehensive evaluation shows that proposed achieves accurate stitching, significantly reduced ghosting effects overlapping regions results other The comparative experiments demonstrate outperforms state-of-the-art warps higher precision less distortion overlapping. algorithm illustrates great application potential which can achieve stitching.