SIA OpenIR  > 机器人学研究室
改进窗口特征及微分算子的立体匹配算法
Alternative TitleStereo Matching by Improved Window Characteristics and Differential Operators
李新春1; 殷新勇1; 林森1,2,3
Department机器人学研究室
Source Publication激光与光电子学进展
ISSN1006-4125
2020
Volume57Issue:8Pages:1-12
Indexed ByCSCD
CSCD IDCSCD:6719066
Contribution Rank2
Funding Organization国家自然科学基金(61473280,61773367) ; 国家自然科学基金委员会“共融机器人基础理论与关键技术研究”重大研究计划(91648118) ; 辽宁省教育厅科学研究一般项目(L2014132) ; 辽宁省自然科学基金面上项目(2015020100)
Keyword机器视觉 立体匹配 特征信息描述 匹配描述子 水下图像
Abstract

ELAS算法是一种性能优良的典型传统立体匹配算法,但该算法视差图条纹明显且具有空洞区域。针对这一问题,提出一种匹配窗口特性与微分特性相结合的局部立体匹配算法,通过增强描述子对点特征信息的描述,为待匹配点提供更有区分度的相似性度量。首先根据彩色图像的经典自适应算法,从空间上提出了适应于灰度图像的窗口描述子,该窗口在待匹配点的邻域内具有固定大小和形状的特征,避免在匹配过程中反复计算匹配窗口。其次依据图像信号的特点,从像素层面上选择平滑性更小的微分算子。然后将提出的匹配窗口与微分算子相结合,获得了比只使用两者之一更强的特性信息描述能力。最后通过标准数据集客观检验与自采集图像主观评价,表明算法更为稳健,而且具有更高的匹配精度,明显改善了原匹配策略视差图中出现的条纹及空洞的问题。

Other Abstract

ELAS algorithm is a typical traditional matching algorithm with good performance, but the disparity maps of the algorithm has obvious stripes and void regions. To resolve these issues, a stereo matching algorithm is proposed in this paper, which combines the characteristics of matching window with the differentials. By enhancing the descriptor's feature information of the points, the discrimination of similarity measure is provided for the matching points. Firstly, according to the classical adaptive algorithm of color images, a window descriptor adapted to the gray image is proposed spatially. The window feature has a fixed size and shape in the field of the matching points, which avoids repeatedly calculating the matching window in the matching process. Next, according to the characteristics of image signal, the less smoothness differential operator is selected at the pixel level. Then the proposed matching window is combined with the differential operator to obtain a stronger description ability of feature information than either of the two. Lastly, through the objective evaluation of standard benchmark and subjective evaluation of self-collected images show that the proposed algorithm is more robust and has higher matching accuracy, and obviously improves the stripes issues and minimizes void regions in the disparity map.

Language中文
Citation statistics
Document Type期刊论文
Identifierhttp://ir.sia.cn/handle/173321/25644
Collection机器人学研究室
Corresponding Author殷新勇
Affiliation1.辽宁工程技术大学电子与信息工程学院
2.中国科学院沈阳自动化研究所机器人学国家重点实验室
3.中国科学院机器人与智能制造创新研究院
Recommended Citation
GB/T 7714
李新春,殷新勇,林森. 改进窗口特征及微分算子的立体匹配算法[J]. 激光与光电子学进展,2020,57(8):1-12.
APA 李新春,殷新勇,&林森.(2020).改进窗口特征及微分算子的立体匹配算法.激光与光电子学进展,57(8),1-12.
MLA 李新春,et al."改进窗口特征及微分算子的立体匹配算法".激光与光电子学进展 57.8(2020):1-12.
Files in This Item:
File Name/Size DocType Version Access License
改进窗口特征及微分算子的立体匹配算法.p(2733KB)期刊论文作者接受稿开放获取CC BY-NC-SAView Application Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[李新春]'s Articles
[殷新勇]'s Articles
[林森]'s Articles
Baidu academic
Similar articles in Baidu academic
[李新春]'s Articles
[殷新勇]'s Articles
[林森]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[李新春]'s Articles
[殷新勇]'s Articles
[林森]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: 改进窗口特征及微分算子的立体匹配算法.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.