采用多尺度方向微分比率的角点检测算法

王富平; 水鹏朗

doi:10.7652/xjtuxb201604011

您当前的位置：

首页 >

文章列表页 >

采用多尺度方向微分比率的角点检测算法

更新时间：2025-07-09

- 采用多尺度方向微分比率的角点检测算法
- A Corner Detection Algorithm Based on Multi-Scale Directional Differential Ratio
- 西安交通大学学报 2016年50卷第4期页码：68-75
- 作者机构：
  
  西安电子科技大学雷达信号处理国家重点实验室,西安,710071
- 作者简介：
- 基金信息：
  
  : 2015-09-10。作者简介: 王富平(1987—),男,博士生;水鹏朗(通信作者),男,教授,博士生导师。基金项目: 国家自然科学基金资助项目(61271295)
- DOI：10.7652/xjtuxb201604011
  中图分类号： TN911.7
- 网络首发：2016-04-10，
  
  纸质出版：2016
- 稿件说明：
移动端阅览
王富平, 水鹏朗. 采用多尺度方向微分比率的角点检测算法[J]. 西安交通大学学报, 2016,50(4):68-75.

A Corner Detection Algorithm Based on Multi-Scale Directional Differential Ratio[J]. 2016, 50(4): 68-75.
王富平, 水鹏朗. 采用多尺度方向微分比率的角点检测算法[J]. 西安交通大学学报, 2016,50(4):68-75. DOI： 10.7652/xjtuxb201604011.

A Corner Detection Algorithm Based on Multi-Scale Directional Differential Ratio[J]. 2016, 50(4): 68-75. DOI： 10.7652/xjtuxb201604011.

摘要

针对相邻图像结构间的相互干扰会导致角点错检率升高的问题

提出了一种基于多尺度方向微分比率(MDDR)的角点检测算法。该算法首先利用Canny边缘检测算法提取原始灰度图像的边缘轮廓; 对于每一个轮廓像素

分别利用各向同性导数和各向异性高斯方向导数滤波器提取主方向及其正交方向上的方向强度微分信息; 然后将3个尺度下微分比率融合为MDDR测度; 最后进行阈值和非极大值抑制处理获得最终角点。不同于传统角点检测算法中只使用单一滤波器

MDDR算法中使用2种不同类型的滤波器

能在精确提取角点附近不同方向微分信息的同时避免邻近图像结构间的相互干扰

从而提高了角点定位准确性并增强了测度的噪声鲁棒性。实验结果表明

MDDR算法的平均检测准确率比点到弦距离累积算法提高了27.1%

并且平均错检率比残余面积算法和Gabor算法分别低28.4%和32.4%。

Abstract

A new corner detection algorithm using multi-scale directional differential ratio(MDDR)of images is proposed to solve the problem that the interference between adjacent image structures may lead to the increase of the false detection ratio. The algorithm firstly extracts the image edge contour of an original grayscale image using the Canny edge detection algorithm. The differential information along the principal direction and its perpendicular direction of each contour pixel is extracted by using the isotropic Gaussian directional derivative filter and the anisotropic Gaussian directional derivative filter

respectively. Then

the differential ratios at three scales are fused into the MDDR corner measure. Final corners are then obtained through thresholding the fused measure followed by the non-maximum suppression. The MDDR algorithm is different from traditional corner detection algorithms that use a single filter

it uses two different kinds of filters to precisely extract the differential information along different directions around the corners and to avoid the interference between adjacent image structures. Thus

the corner localization accuracy is improved

and the noise-robustness of corner measures is enhanced. Experimental results and a comparison with the chord-to-point distance accumulation algorithm show that the average detection accuracy of the proposed algorithm increases by 27.1%

and the average false detection ratio of the proposed algorithm is 28.4% and 32.4% lower than those of the residual area and Gabor detection algorithms

respectively.

关键词

Keywords

references

SMITH S M, BRADY J M. SUSAN-A new approach to low level image processing [J]. International Journal of Computer Vision, 1997, 23(1): 45-78.

ROSTEN E, POTER R, DRUMMOND T. Faster and better: a machine learning approach to corner detection [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(1): 105-119.

LAN J, ZHANG M. Fast and robust corner detector based on double-circle mask [J]. Optical Engineering, 2010, 49(12): 1-8.

王浩, 周祚峰, 曹剑中, 等. 基于边缘轮廓上多尺度自相关矩阵的角点检测 [J]. 系统工程与电子技术, 2014, 36(6): 1220-1224.

WANG Hao, ZHOU Zuofeng, CAO Jianzhong, et al. Corner detection via multi-scale autocorrelation matrix on edge contours [J]. Systems Engineering and Electronics, 2014, 36(6): 1220-1224.

HE X C, YUNG N H C. Corner detector based on global and local curvature properties [J]. Optical Engineering, 2008, 47(5): 1-12.

AWRANGJEB M, LU G J. Robust image corner detection based on the chord-to-point distance accumulation technique [J]. IEEE Transactions on Multimedia, 2008, 10(6): 1059-1072.

TENG S W, SADAT R M, LU G J. Effective and efficient contour-based corner detectors [J]. Pattern Recognition, 2015, 48(7): 2185-2197.

邢远秀, 章登义, 赵俭辉. 利用多尺度弦角尖锐度累积的自适应角点检测算子 [J]. 武汉大学学报: 信息科学版, 2015, 40(5): 617-622.

XING Yuanxiu, ZHAO Dengyi, ZHAO Jianhui. An adaptive threshold corner detector based on multi-scale chord-angle sharpness accumulation [J]. Geomatics and Information Science of Wuhan University, 2015, 40(5): 617-622.

SHUI P L, ZHANG W C. Corner detection and classification using anisotropic directional derivative representations [J]. IEEE Transactions on Image Processing, 2013, 22(8): 3204-3218.

ZHANG W C, WANG F P. Corner detector using Gabor filters [J]. IET Image Processing, 2014, 8(11): 639-646.

章为川, 水鹏朗, 朱磊. 利用各向异性高斯方向导数相关矩阵的角点检测方法 [J]. 西安交通大学学报, 2012, 46(11): 91-97.

ZHANG Weichuan, SHUI Penglang, ZHU Lei. A corner detection algorithm using anisotropic Gaussian directional derivatives autocorrelation matrix on edge contours [J]. Journal of Xi'an Jiaotong University, 2012, 46(11): 91-97.

ZHANG W C, SHUI P L. Contour-based corner detection via angle difference of principal directions of anisotropic Gaussian directional derivatives [J]. Pattern Recognition, 2015, 48(9): 2785-2797.

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

多尺度特征融合的铁轨异物入侵检测研究

一种Roberts自适应边缘检测方法

一种有效抑制睫毛干扰的虹膜定位算法

一种多尺度X射线胸片图像增强算法

利用各向异性高斯方向导数相关矩阵的角点检测方法