VR全景图拼接关键技术研究

打开文本图片集
摘 要:随着5G时代的来临,全景VR技术应用呈现爆发式增长。为解决现有全景拼接软件的不足,提出一种图像拼接改进算法,将SIFT,改进RANSAC及多维索引树结合起来,实验表明,算法拼接的图像清晰度比较好,具有较好的抗噪能力和较好的实用性。
关键词:全景图拼接技术;SIFT算法;RANSAC算法;单应性矩阵;KDG-tree
中图分类号:TP391 文献标志码:A 文章编号:2095-2945(2022)10-0145-04
Abstract: With the advent of the 5G era, the application of panoramic VR technology shows explosive growth. In order to solve the deficiency of existing panoramic mosaic software, an improved image stitching algorithm is proposed, which combines SIFT, improved RANSAC and multi-dimensional index tree. The experimental results show that the image stitching algorithm has satisfying image definition, anti-noise capacity and practicability.
Keywords: panoramic mosaic technology; SIFT algorithm; RANSAC algorithm; homography matrix; KDG-tree
全景图是VR中一种实际场景表示方法,通常要获得高质量的全景图需要购买昂贵的专用拍摄设备,并且设备需要专业人员精确地校准,因此获取全景图成本高。(剩余4603字)