An improved SIFT algorithm for image registration based realization of the vision figure

被引：0

作者：

Meng, Qingsong ^{[1
]}

Lv, Zhihui ^{[1
]}

机构：

[1] School of Automation, Harbin University of Science and Technology, Harbin, China

来源：

International Journal of Control and Automation | 2016年 / 9卷 / 06期

关键词：

Program debugging - Image enhancement;

D O I：

10.14257/ijca.2016.9.6.04

中图分类号：

学科分类号：

摘要：

The purpose of this paper is to eliminate the mismatch points of image registration. Analyzing the SIFT algorithm and finding mismatch source, according to the source to develop solutions to eliminate the error matching points. And compared with the traditional SIFT registration results. In this paper, firstly do initial matching with SIFT algorithm, then use the feature of key points in spatial domain to eliminate the error matching, Finally, using RANSAC algorithm to do the final debugging. Experimental results show that the improved SIFT algorithm can better eliminate mismatch points. © 2016 SERSC.

引用

页码：51 / 58