基于SURF算法的实时拼接研究  

Research on real-time splicing based on the SURF algorithm

作　　者：冯向东[1] FENG Xiangdong(The Engineering and Technical College,Chengdu University of Technology,Leshan 614000,China)

机构地区：[1]成都理工大学工程技术学院,四川乐山614000

出　　处：《电子设计工程》2023年第6期10-14,共5页Electronic Design Engineering

基　　金：四川省科技厅-中央引导地方项目(2021ZYD0021);乐山市科技局科技计划项目(21GZD015);数学地质四川省重点实验室开放基金资助课题(scsxdz2019by01)。

摘　　要：视频拼接有助于人们获得更高分辨率、大视野的场景。文中主要对视频拼接过程中的图像配准及图像融合的SURF(Speeded Up Robust Features)算法进行研究,详细地阐述了基于SURF特征实现图像拼接的每一个步骤,从而快速提高特征点的提取速度,提高了图像配准过程中的精度与准确率,加快了图像配准的速度,并提升图像融合的效果。实验结果表明,该算法能够有效地抑制视差伪影,其拼接图像无明显接缝,并且可以实现实时拼接,拼接的图像更加自然,拼接的视觉效果较好。Video splicing can help people to obtain scenes with higher resolution and wide field of vision.This paper mainly studies the surf algorithm of video image registration and video image fusion in video splicing. Each step of image mosaic based on surf feature is described in detail,by trying to quickly improve the extraction speed of feature points,and improve the accuracy and precision of registration,and further accelerate the speed of registration and the effect of image fusion. Experimental results show that the algorithm can effectively suppress parallax artifacts,and the stitched images have not obvious seams. And it can realize real-time splicing,the spliced image is more natural,and the visual effect of splicing is better.

关 键 词：SURF算法 视频拼接 特征点提取 图像融合 图像配准 

分 类 号：TN957.52[电子电信—信号与信息处理]