全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...

对数极坐标系下尺度不变特征点的检测与描述

DOI: 10.11834/jig.20151209

Keywords: 计算机视觉,图像匹配,对数极坐标系,尺度不变特征点,描述符

Full-Text   Cite this paper   Add to My Lib

Abstract:

目的当前国际流行的SIFT算法及其改进算法在检测与描述特征点时基于高斯差分函数,存在损失图像高频信息的缺陷,从而导致图像匹配时其性能随着图像变形的增加而出现急剧下降。针对SIFT算法及其改进算法的这一缺陷,本研究提出了一种新的无图像信息损失的、在对数极坐标系下的尺度不变特征点检测与描述算法。方法本研究提出的尺度不变特征点检测与描述算法首先将直角坐标系下以采样点为中心的圆形图块转换为对数极坐标系下的矩形图块,并以此矩形图块为基础对采样点进行特征点检测与描述符提取;该算法使用固定宽度的窗口在采样点的对数极坐标径向梯度图像的logtr轴上进行移动以判断该点是否为特征点并计算该点的特征尺度,并在具有局部极大窗口响应的特征尺度位置处提取特征点的描述符。该算法的描述符基于对数极坐标系下的矩形图块的灰度梯度的幅值与角度,是一个192维向量,并具有对于尺度、旋转、光照等变化的不变性。结果本研究采用INRIA数据组和Mikolajczyk提出的匹配性能指标对SIFT算法、SURF算法和提出的尺度不变特征点检测与描述算法进行比较。与SIFT算法和SURF算法相比,提出的尺度不变特征点检测与描述算法在对应点数、重复率、正确匹配点数和匹配率等方面均具有一定优势。结论提出了一种基于对数极坐标系的图像匹配算法,即将直角坐标系下以采样点为中心的圆形图块转换为对数极坐标系下的矩形图块,这样在特征点的检测过程中,可以有效规避SIFT算法因为采用DoG函数而造成的高频信息损失;在描述符提取过程中,对数极坐标系可以有效地减少图像的变化量,从而提高了匹配性能。

References

[1]  Wang H L, Jia W B. Literature review on algorithms of image matching[C]//Progress of Computer Technology and Application. Hefei:Press of USCT, 2008:418-423.[王宏力,贾万波.图像匹配算法研究综述[C]//计算机技术与应用进展.合肥:中国科学技术大学出版社, 2008:418-423.]
[2]  Zhang Y, Chu C H. One-dimensional mapping for estimating projective transformations[C]//IEEE Trans. on Image Process. New York, United States of America:IEEE, 2010, 19(11):3049-3058.[DOI:10.1109/TIP.2010.2050724]
[3]  Pan Z A. Research on image matching based on the SIFT algorithm[D]. Xi\'an:Xidian University, 2012.[潘子昂.基于SIFT算法的图像匹配研究[D].西安:西安电子科技大学, 2012.]
[4]  Zhao Q. Research on algorithm of image matching[D]. Xi\'an Xidian University, 2013.[赵启.图像匹配算法研究[D].西安:西安电子科技大学, 2013.]
[5]  Barnea D I, Silverman H F. A class of algorithms for fast digital image registration[C]//IEEE Transactions of Computing. New York, United States of America:IEEE, 1972:179-186.[DOI:10.1109/TC.1972.5008923]
[6]  Dai T, Zhu C R, Hu S P. Literature review on technologies of image matching[J]. Digital Technology and Application, 2012:174-177.[戴涛,朱长仁,胡树平.图像匹配技术综述[J].数字技术与应用, 2012:174-177.]
[7]  Moravec H. Obstacle avoidance and navigation in the real world by a seeing robot rover[R]//Tech Report CMU-RI-TR-3. United States of America:Carnegie-Mellon University, Robotics Institute, 1980.
[8]  Zokai S, Wolberg G. Image registration using log-polar mappings for recovery of large-scale similarity and projective transformation[C]//Proceedings of IEEE Trans. on Image Process. New York, United States of America:IEEE, 2005, 14(10):1422-1434.[DOI:10.1109/TIP. 2005. 854501]
[9]  Mikolajczyk K, Tuytelaars T, Schmid C, et al. A comparison of affine region detectors[C]//Proceedings of Int\'l J. Computer Vision. Netherlands:Kluwer Academic Publishers, 2005, 65(1-2):43-72.[DOI:10.1007/s11263-005-3848-x]
[10]  Mikolajczyk K, Schmid C. Scale and affine invariant interest point detectors[C]//Proceedings of Int\'l J. Computer Vision. Netherlands:Kluwer Academic Publishers, 2004, 60(1):63-86.[DOI:10.1023/B:VISI.0000027790.02288.f2]
[11]  Shen S Z, Zhang X L, Heng W. Improved Harris corner detection algorithm based on auto-adaptive threshold and pre-selection[J]. Journal of Data Acquisition & Processing, 2011, 26(2):207-213.[沈世?,张小龙,衡伟.一种自适应阈值的预筛选Harris角点检测方法[J].数据采集与处理, 2011, 26(2):207-213.] [DOI:10.3969/j.issn.1004-9037. 2011. 02.016]
[12]  Chen B P, Zhao J L, Yin Z L. A fast approach to realize bilinear interpolation algorithm[J]. Journal of Beijing Electronic Science and Technology Institute, 2004, 12(4):21-23.[陈宝平,赵俊岚,尹志凌.双线性插值算法的一种快速实现方式[J].北京电子科技学院学报, 2004, 12(4):21-23.] [DOI:10.3969/j.issn.1672-464X.2004.04.006]
[13]  Mikolajczyk K, Schmid C. A performance evaluation of local descriptors[C]//Proceedings of IEEE Trans. on Pattern Analysis and Machine Intelligence. New York, United States of America:IEEE, 2005, 27(10):1615-1630.[DOI:10.1109/TPAMI.2005.188]
[14]  Wang Q S, Zhao X A, Lu J G, et al. Feature point extraction with improved Harris algorithm based on difference of Gaussian[J]. Science of Surveying and Mapping, 2014, 39(4):119-134.[王青松,赵西安,吕京国,等.基于高斯差分的改进Harris特征点提取算法[J].测绘科学, 2014, 39(4):119-134.]
[15]  Zhang J, Sang H S. Parallel architecture for DoG scale-spcae construction[J]. Microelectronics & Computer, 2014, 31(7):6-9.[张静,桑红石.构造高斯差分尺度空间的并行结构[J].微电子学与计算机, 2014, 31(7):6-9.]
[16]  Rosenfeld A, Kak A C. Digital picture processing vol. I and Ⅱ[M]. Orlando, FL:Academic Press, 1982:127-138.
[17]  Harris C, Stephens M. A combined corner and edge detector[C]//Proc. of 4th Alvey Vision Conference. Manchester, United Kingdom:The Plessey Company pic, 1988:147-151.
[18]  Lowe D. Distinctive image features from scale-invariant keypoints[C]//Int\'l J. Computer Vision. Netherlands:Kluwer Academic Publishers, 2004, 60(2):91-110.[DOI:10.1023/B:VISI.0000029664.99615.94]
[19]  Bay H, Tuytelaars T, van Gool L. SURF:speeded up robust features[C]//Proceedings of the 9th European Conference on Computer Vision. Berlin Heidelberg:Springer, 2006:404-417.[DOI:10.1007/11744023_32]

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133