一种改进的Harris角点检测算法
|
摘要
针对Harris算法存在运算慢、抗噪能力差以及在实际应用中存在不必要角点簇等问题,提出了一种改进的Harris角点检测算法。采用加速分割测试特征点检测原理,排除大量的非特征点得到初始点,以初始点响应Harris函数执行非极大值抑制,保留局部角点响应函数最大值的像素点,以这些点为中心,以一定半径搜索角点簇,采用容忍距离内保留一个特征点,降低角点簇影响。提取Harris角点后,采用NCC算法进行粗匹配,再用RANSAC算法消除误匹配,提高图像拼接的精度。实验结果表明,该算法能提高检测的速度,去除大量伪角点和不必要角点簇,验证了改进算法的有效性和实用性。
Harris algorithm exist the problems of high computational cost,poor noise immunity,and the presence of unnecessary corner cluster in the practical applications,so an improved Harris corner detection algorithm is proposed.The method uses the features from accelerated segment test to exclude a large number of non-feature points.Then the detected points as the initial point response function Harris performs non-maxima suppression,retain the local corner response function of maximum pixels point,and take these points as the center,in a certain corner radius search cluster,within tolerance is just only one feature point to reduce the influence of the angular point cluster.In order to improve the accuracy of image mosaic,after Harris corner is extracted,the coarse matching is done by NCC algorithm,then RANSAC algorithm is used to eliminate the false matching.The experimental results show that the improved method increases the speed of detection,removes a large number of pseudo corner points and unnecessary corner clusters.The effectiveness and feasibility of the proposed method is verified.
Harris algorithm exist the problems of high computational cost,poor noise immunity,and the presence of unnecessary corner cluster in the practical applications,so an improved Harris corner detection algorithm is proposed.The method uses the features from accelerated segment test to exclude a large number of non-feature points.Then the detected points as the initial point response function Harris performs non-maxima suppression,retain the local corner response function of maximum pixels point,and take these points as the center,in a certain corner radius search cluster,within tolerance is just only one feature point to reduce the influence of the angular point cluster.In order to improve the accuracy of image mosaic,after Harris corner is extracted,the coarse matching is done by NCC algorithm,then RANSAC algorithm is used to eliminate the false matching.The experimental results show that the improved method increases the speed of detection,removes a large number of pseudo corner points and unnecessary corner clusters.The effectiveness and feasibility of the proposed method is verified.
引文
[1]何海清,黄声享.改进的Harris亚像素角点快速定位[J].中国图象图形学报,2012,17(7):853-857.
[2]王民,周兆镇,李昌华,等.基于像素点灰度差的Harri角点检测算法[J].计算机工程,2015,41(6):227-230.
[3]LOWE D G.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.
[4]XU Jian,CHEN Xiaoyuan,SONG Xiaoping,et al.An improved Harris-FAST algorithm for underwater objectcorner detection[C]//IEEE 27th Chinese Control and Decision Conference,2015:5424-5418.
[5]于来行,冯林,张晶,等.自适应融合目标和背景的图像特征提取方法[J].计算机辅助设计与图形学学报,2016,28(8):1250-1259.
[6]FARZIN M,RIKU S.Robust image corner detection through curvature scalespace[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1998,20(12):1376-1381.
[7]MORAVEC H P.Towards automatic visual obstacle avoidance[C]//Proceedings of International Joint Conference on Artificial Intelligence,1977:584-590.
[8]HARRIS C,STEPHENS M.A combined corner and edge detector[C]//Proceedings of the 4th Alvey Vision Conference,Manchester,1988:147-151.
[9]SMITH S M,BRADY M.A new approach to low level image processing[J].International Journal of Computer Version,1997,23(1):45-78.
[10]扈立超,史再峰,庞科,等.用于图像匹配的改进Harris特征点检测算法[J].计算机工程,2015,41(10):216-220.
[11]SUN Liang,WANG Shuangqiu,XING Jianchun,et al.Self-adaption Harris corner detection algorithm based on image contrast area[C]//IEEE 27th Chinese Control and Decision Conference,2015:2287-2290.
[12]ROSTEN E,DRUMMOND T.Faster and better:a machine learning approach to comer detection[J].IEEE Transcations on Pattern Analysis and Machine Intelligence,2010,32(1):105-119.
[2]王民,周兆镇,李昌华,等.基于像素点灰度差的Harri角点检测算法[J].计算机工程,2015,41(6):227-230.
[3]LOWE D G.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.
[4]XU Jian,CHEN Xiaoyuan,SONG Xiaoping,et al.An improved Harris-FAST algorithm for underwater objectcorner detection[C]//IEEE 27th Chinese Control and Decision Conference,2015:5424-5418.
[5]于来行,冯林,张晶,等.自适应融合目标和背景的图像特征提取方法[J].计算机辅助设计与图形学学报,2016,28(8):1250-1259.
[6]FARZIN M,RIKU S.Robust image corner detection through curvature scalespace[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1998,20(12):1376-1381.
[7]MORAVEC H P.Towards automatic visual obstacle avoidance[C]//Proceedings of International Joint Conference on Artificial Intelligence,1977:584-590.
[8]HARRIS C,STEPHENS M.A combined corner and edge detector[C]//Proceedings of the 4th Alvey Vision Conference,Manchester,1988:147-151.
[9]SMITH S M,BRADY M.A new approach to low level image processing[J].International Journal of Computer Version,1997,23(1):45-78.
[10]扈立超,史再峰,庞科,等.用于图像匹配的改进Harris特征点检测算法[J].计算机工程,2015,41(10):216-220.
[11]SUN Liang,WANG Shuangqiu,XING Jianchun,et al.Self-adaption Harris corner detection algorithm based on image contrast area[C]//IEEE 27th Chinese Control and Decision Conference,2015:2287-2290.
[12]ROSTEN E,DRUMMOND T.Faster and better:a machine learning approach to comer detection[J].IEEE Transcations on Pattern Analysis and Machine Intelligence,2010,32(1):105-119.