Smooth Stitching Method of VR Panoramic Image Based on Improved SIFT Algorithm
Qiaohui Liu · 2022
Aiming at the shortcomings of the traditional Harris algorithm, a smooth mosaic square of VR panoramic image based on the Improved SIFT algorithm is proposed. Firstly, the SIFT feature points are extracted by the SIFT detection algorithm, and then the Harris corners of VR panoramic image are detected in the neighborhood of the SIFT feature points to enhance the smoothness of the feature points of the panoramic image. At the same time, with the help of the invariance of SIFT features to rotation and scale, as well as the good smoothness to noise interference and brightness change, the problem that VR panoramic image mosaic is sensitive to scale change is solved, which makes the smooth mosaic of VR panoramic image possible. The experimental results show that the smooth stitching method of VR panoramic image based on the Improved SIFT algorithm has good smoothness, and the matching speed is also significantly improved.