摘要
针对标准SIFT算法在虚拟场景绘制的图像拼接中还存在拼接效果不好等问题。本文提出了一种基于拼接优化SIFT算法的虚拟场景绘制图像拼接模型。首先采用高斯函数构建尺度空间,然后使用最小欧式距离进行初拼接,接着利用投影变换模型进行特征点的"外点"去除,最后通过提取的特征点进行图像的拼接。通过仿真实验表明,通过对标准SIFT算法进行双层拼接优化后,改进算法相比较标准算法具有更好的图像拼接效果。
According to the standard SIFT algorithm in the virtual scene rendering image stitching isjoining together the result is bad.This paper proposes a virtual scene rendering based on splicing optimalSIFT algorithm image matching model.Gaussian function was used to construct space first,and thenUSES the minimum Euclidean distance for early,then using the projection transformation model offeature points outside the"point",finally by extracting the characteristic points of image stitching.Through the simulation experiment shows that through the standard double stitching optimized SIFTalgorithm,improved algorithm compared with standard algorithm has better effect of image stitching.
作者
余妹兰
叶群
鲍方云
Yu Meilan;Ye Qun;Bao Fangyun(Hunan Vocational Institute of Safety Technology,Changsha 410151,China;Hangzhou XueJun Investment Management CO.,LTD,Chairman of the Board of Directors,Hangzhou 310007,China)
出处
《科技通报》
北大核心
2017年第7期133-136,共4页
Bulletin of Science and Technology
基金
湖南省科技厅重点研发计划项目(编号:2015ZK3071)
国家安全生产监督管理总局安全生产重大事故防治关键技术科技项目(编号:hunan-0016-2016AQ)