|本期目录/Table of Contents|

 基于加权引导滤波的局部立体匹配算法(PDF)

《纳米技术与精密工程》[ISSN:1672-6030/CN:12-1351/O3]

期数:
2017年5期
页码:
394-399
栏目:
精密测量
出版日期:
2017-09-15

文章信息/Info

Title:
 Local Stereo Matching Algorithm Based on Weighted Guided Filter
文章编号:
1672-6030(2017)05-0394-06
作者:
 陈松 陈晓冬 苏修 刘依林 汪毅 郁道银
 (天津大学光电信息技术教育部重点实验室, 天津 300072)
Author(s):
Chen Song Chen Xiaodong Su Xiu Liu Yilin Wang Yi Yu Daoyin
 (Key Laboratory of Opto-Electronic Information Technology of Ministry of Education, Tianjin University, Tianjin 300072, China)
关键词:
 双目立体视觉 立体匹配 加权引导滤波 自适应窗口
Keywords:
 binocular stereo vision stereo matching weighted guided filter adaptive window
分类号:
TN911.73
DOI:
10.13494/j.npe.20160053
文献标识码:
A
摘要:
 立体匹配是双目立体视觉中的关键环节.本文在现有局部立体匹配算法的基础上,提出一种基于加权引导滤波的局部立体匹配算法.该算法在代价聚合阶段采用加权引导滤波方法,根据不同窗口像素纹理的丰富程度,对引导滤波的规整化参数进行自适应调整,实现更准确的代价聚合;在视差选取阶段,根据聚合后的代价空间信息对视差的可靠性进行判断,对于视差不可靠像素,利用自适应窗口方法进一步代价聚合后确定视差.在MATLAB软件平台上利用标准立体图像对进行测试,实验结果表明,本文提出的算法平均误匹配率为5.20%,相比于现有的局部立体匹配算法具有更高的精度.
Abstract:
 Stereo matching is a key link in binocular stereo vision. With the existing local stereo matching algorithms, this paper proposes a local stereo matching algorithm based on weighted guided filter. This algorithm uses weighted guided filter in the cost aggregation phase and the uniform parameter of guided filter adjustments adaptively according to the abundance of texture in different windows to achieve a more accurate cost aggregation. In the disparity selection phase, the algorithm judges the reliability of the disparity according to the aggregated cost volume and determines the disparity after further cost aggregation using the adaptive window method for the disparity unreliable pixels. Standard stereo image pairs are used for testing on MATLAB. Experimental results show that the average matching error rate of the algorithm proposed in this paper is 5.20% and this algorithm has higher accuracy compared with existing local stereo matching algorithms.

参考文献/References

备注/Memo

备注/Memo:
收稿日期: 2017-04-06.
基金项目: 国家自然科学基金仪器专项资助项目(61327007).
作者简介: 陈松(1992— ),男,硕士研究生.
通讯作者: 陈晓冬,教授,xdchen@tju.edu.cn.
更新日期/Last Update: 2017-09-28