采用黎曼度量的Hausdorff距离及其在图像匹配中的应用
Hausdorff distance based on Riemannian metric and its application in image matching
查看参考文献15篇
文摘
|
现有的基于Hausdorff距离的边缘图像匹配利用边缘的位置信息, 忽视了边缘的其他有用信息. 为了提高基于边缘的图像匹配的鲁棒性, 提出了一种基于黎曼度量的Hausdorff距离(RM -HD)图像匹配算法. 通过边缘点的灰度和附近梯度信息构造了边缘结构张量, 由于结构张量具有流形结构, 采用流形的测地距离来度量边缘结构张量的距离, 用其对Hausdorff距离进行加权, 设计了图像匹配算法. 分别用可见光和红外图像进行实验, 结果表明: RM -HD算法在匹配准确性、抵抗噪声干扰、光照变化等方面性能良好 |
其他语种文摘
|
The existing Hausdorff distance based image matching method mainly uses the location of edges, ignoring other useful information of edge. In order to improve the robustness of edge-based image matching, a new method of image matching based on Hausdorff distance of Riemannian metric (RM-HD) was proposed. The edge structure tensor was constructed by intensity feature and edge gradient, since the structure tensor possessed a Riemannian manifold, the distance between structure tensors was measured by the geodesic distance on Riemannian manifold. Then, the HD was weighted by the structure tensors. Finally, a image matching algorithm based on RM-HD was designed. Experimental results on visible images and infrared image show that the proposed RM-HD possesses good performance in terms of matching accuracy, robustness to random noise and illumination |
来源
|
红外与激光工程
,2011,40(2):365-369 【核心库】
|
关键词
|
图像匹配
;
Hausdorff距离
;
边缘结构
;
结构张量
;
黎曼度量
|
地址
|
中国科学院沈阳自动化研究所, 辽宁, 沈阳, 110016
|
语种
|
中文 |
文献类型
|
研究性论文 |
ISSN
|
1007-2276 |
学科
|
自动化技术、计算机技术 |
基金
|
中国科学院国防科技创新基金
|
文献收藏号
|
CSCD:4147204
|
参考文献 共
15
共1页
|
1.
Huttenlocher D P. Comparing images using the Hausdorff distance.
IEEE Transaction on Pattern Analysis and Machine Intelligence,1993,15(9):850-863
|
被引
300
次
|
|
|
|
2.
Dubuisson M P. A modified Hausdorff distance for object matching.
Proceedings of International Conference on Pattern Recognition,1994:566-568
|
被引
6
次
|
|
|
|
3.
Sim DG. Object matching algorithms using robust Hausdorff distance measures.
IEEE Transaction on Image Processing,1999,8(3):425-429
|
被引
70
次
|
|
|
|
4.
Zhao C J. A new Hausdorff distance for image matching.
Pattern Recognition Letter,2005,26(5):581-586
|
被引
8
次
|
|
|
|
5.
Park S C. Tracking nonrigid objects using probabilistic Hausdorff distance matching.
Pattern Recognition,2005,38(12):2373-2384
|
被引
2
次
|
|
|
|
6.
Vivek E P. Robust Hausdorff distance measure for face recognition.
Pattern Recognition,2007,40(2):431-442
|
被引
8
次
|
|
|
|
7.
Yoruk E. Shape-based hand recognition.
IEEE Transaction on Image Processing,2006,15(7):1803-1825
|
被引
10
次
|
|
|
|
8.
张志佳. 基于区域模板相关的图像匹配方法研究.
红外与激光工程,2004,33(6):615-618
|
被引
2
次
|
|
|
|
9.
杨兵. 基于新的加权Hausdorff距离的图像匹配.
红外与激光工程,2007,36(5):7515-7754
|
被引
1
次
|
|
|
|
10.
Harris C. A combined comer and edge detector.
Alvey Conference,1988:147-152
|
被引
2
次
|
|
|
|
11.
Wang Haiyun. Accurate optical flow estimation using adaptive scale-space and 3D structure tensor.
ICIP(2),2002:301-304
|
被引
1
次
|
|
|
|
12.
Pennec X. A riemannian framework for tensor computing.
International Journal of Computer Vision,2006,66(1):41-66
|
被引
50
次
|
|
|
|
13.
Tuzel O. Human detection via classification on riemannian manifolds.
CVPR,2007:1-8
|
被引
1
次
|
|
|
|
14.
Tuzel O. Region covariance: a fast descriptor for detection and classification.
Proceeding of 9th ECCV,2006:589-600
|
被引
1
次
|
|
|
|
15.
Bartoli A. Groupwise geometric and photometric direct image registration.
IEEE Transaction on Pattern Analysis and Machine Intelligence,2008,30(12):2098-2108
|
被引
6
次
|
|
|
|
|