局部不变特征匹配的并行加速技术研究

王灿进; 孙涛; 陈娟

doi:10.3788/YJYXS20142902.0266

您当前的位置：

首页 >

文章列表页 >

局部不变特征匹配的并行加速技术研究

成像技术与图像处理 | 更新时间：2020-08-12

- 局部不变特征匹配的并行加速技术研究
- Speeding up local invariant feature matching using parallel technology
- 液晶与显示 2014年29卷第2期页码：266-274
- 作者机构：
  
  1. 中国科学院大学北京,中国,100049
  2. 中国科学院长春光学精密机械与物理研究所激光与物质相互作用国家重点实验室,吉林长春,130033
- 作者简介：
- 基金信息：
  
  院地合作（长吉图专项基金）（No.2011CJT0006）();吉林科技发展计划项目（No.20126015）()
- DOI：10.3788/YJYXS20142902.0266
  中图分类号： TP391.4
- 收稿日期：2013-02-22，
  
  修回日期：2013-04-11，
  
  纸质出版日期：2014-03-05
- 稿件说明：
移动端阅览
王灿进, 孙涛, 陈娟. 局部不变特征匹配的并行加速技术研究[J]. 液晶与显示, 2014,29(2): 266-274

WANG Can-jin, SUN Tao, CHEN Juan. Speeding up local invariant feature matching using parallel technology[J]. , 2014,29(2): 266-274
王灿进, 孙涛, 陈娟. 局部不变特征匹配的并行加速技术研究[J]. 液晶与显示, 2014,29(2): 266-274 DOI： 10.3788/YJYXS20142902.0266.

WANG Can-jin, SUN Tao, CHEN Juan. Speeding up local invariant feature matching using parallel technology[J]. , 2014,29(2): 266-274 DOI： 10.3788/YJYXS20142902.0266.

摘要

针对SIFT、SURF等局部不变特征在大尺寸图像上匹配时过于耗时的问题，将FREAK算子应用于图像匹配中，并提出一种多线程并行加速方法。首先介绍FREAK描述子的特征点的检测、特征描述向量的生成和特征向量的匹配的过程，并分析其优势。其次提出并行处理的2种思路：一是对待匹配图像进行有重叠的分块，对于每一块子图像，开辟新的线程分别进行处理；二是对匹配过程的3个步骤，采用流水线技术进行并行处理，每检测出一个特征点，随即提取出该点的特征向量，然后和模板图像的特征向量集进行匹配。改写SIFT、SURF和FREAK算法进行实验验证，结果证明FREAK计算过程比SIFT和SURF快得多，而并行方法可以在保证匹配精度的同时明显缩短匹配时间。

Abstract

Taking into account the problem that local invariant feature algorithms such as SIFT

SURF take too much time in large images registration

a fast descriptor named FREAK is introduced

and a speeding up method using multi-thread technology is proposed. The process of feature point detection

descriptive characteristics vector generation and feature vector matching of FREAK are introduced and their advantages are analyzed. Then

two ideas of parallel processing are proposed.One is to divide large image into overlapping sub-blocks

and for each sub-image

a new thread is opened up. The other is to use pipelining technology for the three step of registration. Once a feature point is detected

its feature vector is extracted immediately

and soon it matches with the feature vector set of the template image. Rewriting SIFT

SURF and FREAK for experiment

the experimental result indicates that FREAK is faster than the other two algorithms

and our parallel method can greatly economize matching time as well as guarantee the accuracy.

关键词

Keywords

references

Mikolajczyk K, Schmid C.Indexing based on scale invariant interest points[C]//8th IEEE International Conference on Computer Vision, Montbonnot, France:ICCV, 2001:525-531.[2] Tuytelaars T, Gool L. Matching widely separated views based on affine invariant regions[J].International Journal of Computer Vision, 2004, 59(1):61-85.[3] 刘向增, 田铮, 史振广, 等. 基于FKICA-SIFT特征的合成孔径图像多尺度配准[J]. 光学精密工程, 2011, 19(9):2186-2196. Liu X Z, Tian Z, Shi Z G, et al. SAR image multi-scale registration based on FKICA-SIFT features[J].Opt.Precision Eng., 2011, 19(9):2186-2196.(in Chinese)[4] 赵立荣, 朱玮, 曹永刚, 等. 改进的加速鲁棒特征算法在特征匹配中的应用[J]. 光学精密工程, 2013, 21(12):3263-3271. Zhao L R, Zhu W, Cao Y G, et al.Application of improved SURF algorithm to feature matching[J].Opt.Precision Eng., 2013, 21(12): 3263-3271.(in Chinese)[5] 苏可心, 韩广良, 孙海江. 基于 SURF 的抗视角变换图像匹配算法[J]. 液晶与显示, 2013, 28(4): 626-632. Su K X, Han G L, Sun H J.Anti-Viepoint Changing image matching algorithm based on SURF[J].Chinese Journal of Liquid Crystals and Displays, 2013, 28(4): 626-632.(in Chinese)[6] 贾平, 徐宁, 张叶. 基于局部特征提取的目标自动识别[J]. 光学精密工程, 2013, 21(7): 1898-1905. Jia P, Xu N, Zhang Y. Automatic target recognition based on local feature extraction[J].Opt.Precision Eng., 2013, 21(7): 1898-1905.(in Chinese)[7] 李英, 李静宇, 徐正平. 结合SURF与聚类分析方法实现运动目标的快速跟踪[J]. 液晶与显示, 2011, 26(4): 544-550. Li Y, Li J Y, Xu Z P. Moving target fast tracking using SURF and cluster analysis method[J].Chinese Journal of Liquid Crystals and Displays, 2011, 26(4): 544-550. (in Chinese)[8] Lowe D. Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision, 2004, 60(2): 91-110.[9] Brown M, Lowe D. Automatic panoramic image stitching using invariant features[J].International Journal of Computer Vision, 2007, 74(1):59-73.[10] 翟优, 曾峦, 熊伟. 基于不变特征描述符实现星点匹配[J]. 光学精密工程, 2013, 20(11): 2531-2539. Zhai Y, Zeng L, Xiong W. Star matching based on invariant feature descriptor[J].Opt. Precision Eng., 2013, 20(11): 2531-2539.(in Chinese)[11] 吴威, 许廷发, 王亚伟, 等. 高精度全景补偿电子稳像[J]. 中国光学, 2013, 6(3): 378-385. Wu W, X T F, Wang Y W et al.High precision digital image stabilization with full frame compensation[J].Chinese Optics, 2013, 6(3): 378-385.[12] Noah S, Steven M, Richard S. Photo tourism:exploring photo collections in 3D[J].ACM Trans on Graphics, 2006, 25(3):835-846.[13] Bay H, Tuvtellars T, Gool V. SURF: speeded up robust features[C]//Proceedings of the European Conference on Computer Vision, Graz, Austria: ECCV, 2006: 404-417.[14] Calander M, Lepetit V, Strecha C et al. BRIEF: Binary robust Independent elementary features[C]//European Conference on Computer Vision, Crete, Greece: ECCV, 2010: 778-792.[15] Stefan L, Margarita C, Roland S. BRISK: binary robust invariant scalable keypoints[C]//Proceedings of the IEEE International Conference on Computer Vision, Barcelona, Spain: ICCV, 2011: 2548-2555.[16] Rublee E, Rabaud V, Konolige K, et al. Orb: an efficient alternative to sift or surf[C]//Proceedings of the IEEE International Conference on Computer Vision, Barcelona, Spain: ICCV, 2011: 2564-2571.[17] Alahi A, Ortiz R, Vandergheynst P. FREAK:FAST Retina keypoint[C]//Computer Version and Pattern Recognition, Providence, RI, USA: CVPR, 2011: 510-517.[18] Rosten E, Drummond T. Machine learning for highspeed corner detection[C]//9th European Conference on Computer Vision, Graz, Austria: ECCV, 2006: 430-443.[19] Field D, Gauthier J, Sher A, et al.Functional connectivity in the retina at the resolution of photoreceptors[J].Nature, 2010, 467(7316):673-677.[20] 龚卫国, 张璇, 李正浩. 基于改进局部敏感散列算法的图像匹配[J]. 光学精密工程, 2011, 19(16):1375-1383. Gong W G, Zhang X, Li Z H. Image registration based on extended LSH[J].Opt. Precision Eng., 2011, 19(6):1375-1383.(in Chinese)

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于量子克隆多宇宙算法的图像融合研究