基于计算集成成像的图像配准算法的研究.pdfVIP

第37 卷 第5 期 红 外 技 术 Vol.37 No.5 2015 年5 月 Infrared Technology May 2015 〈图像处理与仿真〉 基于计算集成成像的图像配准算法研究 韶阿俊，钱惟贤，顾国华，李 超，毛 晨 （南京理工大学电子工程与光电技术学院，江苏南京 2100 14 ） 摘要：计算集成成像是目前三维重构与显示领域的主流研究方向，是一种基于透镜阵列来获取和重 现目标场景三维的技术，针对集成成像采集的子图特征点匹配问题，传统的图像匹配方法匹配点少、 耗费时间长、容易产生局部极值等问题。为了获取更多的匹配点，在原有 SIFT 配准算法基础上进 行了适当的改进，实现集成成像稠密的图像配准，较大地缩短了配准时间。 关键词：集成成像；三维重构；图像匹配；SIFT 中图分类号：TP391 文献标识码：A 文章编号：1001-8891(2015)05-0398-06 Image Registration Algorithm Research Based on Computational Integral Imaging SHAO A-jun ，QIAN Wei-xian ，GU Guo-hua ，LI Chao，MAO Chen (School of Electronic and Optical Engineering, Nanjing University of Science and Technology , Nanjing 2100 14, China) Abstract ：Computational Integral Imaging （CII ）technology is one of the main research directions in the field of 3D reconstruction and display. It is one kind of technologies based on lens array which is to obtain and reconstruct the goal 3D scene. According to feature point matching problem in integral imaging sub-graphs, the traditional method of image matching is not good enough because of poor matching points, huge time consumption in gand producing local minima. In order to get more matching points, we make the appropriate improvement in original image registration algorithm based on SIFT. It realizes dense image registration and shortens the time consumption. Keywords：computational integral imaging ，3D reconstruction，image registration ，SIFT [3]