基于FAST和BRIEF的图像匹配算法
|
摘要
针对图像匹配算法中匹配率不高以及运算速度较慢等问题,采用改进的FAST(加速分割测试特征)和BRIEF(二进制稳健基元独立特征)算法对图像进行匹配。使用FAST算法提取图像特征点,简化测试模板以提高检测速度;以提取的特征点为中心,使用强度质心方法计算图像块的主方向,根据主方向旋转BRIEF描述器,使其具备旋转不变性;使用易于计算的汉明距离度量各描述器的相似度,据此进行匹配特征。通过和其余算法进行对比实验,验证了该算法在保持高匹配率的同时,降低了计算复杂性。
To solve the problem of low matching rate and poor efficiency of usual image matching algorithm,improved FAST(features from accelerated segment test)and BRIEF(binary robust independent elementary features)were adopted to match images.Firstly,FAST algorithm was used to extract keypoints from images,and test model was simplified to improve detection speed.Secondly,the patch orientation of keypoints was computed using intensity centroid method,and rotation invariance was guaranteed by steering BRIEF descriptors according to the orientation of keypoints.Finally,the keypoints descriptor similarity was evaluated using the Hamming distance,which was very efficient to compute and used to features matching.By comparing with others algorithms,the experimental results show that the algorithm reduces computing complexity while maintaining a relatively good matching rate.
To solve the problem of low matching rate and poor efficiency of usual image matching algorithm,improved FAST(features from accelerated segment test)and BRIEF(binary robust independent elementary features)were adopted to match images.Firstly,FAST algorithm was used to extract keypoints from images,and test model was simplified to improve detection speed.Secondly,the patch orientation of keypoints was computed using intensity centroid method,and rotation invariance was guaranteed by steering BRIEF descriptors according to the orientation of keypoints.Finally,the keypoints descriptor similarity was evaluated using the Hamming distance,which was very efficient to compute and used to features matching.By comparing with others algorithms,the experimental results show that the algorithm reduces computing complexity while maintaining a relatively good matching rate.
引文
[1]FU Lisi,LIU Pengwei,LI Dandan.Improved moment invariant characteristics and object recognition[J].Computer Engineering and Applications,2012,48(31):183-185(in Chinese).[付立思,刘朋维,李丹丹.一种改进的不变矩特征与物体识别[J].计算机工程与应用,2012,48(31):183-185.]
[2]HU Qiong,QIN Lei,HUANG Qingming.A survey on visual human action recognition[J].Chinese Journal of Computers,2013,36(12):2512-2524(in Chinese).[胡琼,秦磊,黄庆明.基于视觉的人体动作识别综述[J].计算机学报,2013,36(12):2512-2524.]
[3]FAN bo,YANG Xiaomei,HU Xueshu.Super-resolution image reconstruction algorithms based on compressive sensing[J].Journal of Computer Applications,2013,33(2):480-483(in Chinese).[樊博,杨晓梅,胡学姝.基于压缩感知的超分辨率图像重建[J].计算机应用,2013,33(2):480-483.]
[4]ZHANG Mingjie,KANG Baosheng.Modified object detection and automatic tracking method[J].Computer Engineering and Design,2014,35(4):1308-1311(in Chinese).[张明杰,康宝生.改进的目标检测与自动跟踪方法研究[J].计算机工程与设计,2014,35(4):1308-1311.]
[5]JIAO Lilong,HAN Xie,LI Dingzhu.Improved image mosaic algorithm based on feature points matching[J].Computer Engineering and Design,2014,35(3):918-922(in Chinese).[焦丽龙,韩燮,李定主.改进的基于特征点匹配的图像拼接融合算法[J].计算机工程与设计,2014,35(3):918-922.]
[6]ZHANG Yong,JI Dongsheng.Improved Harris feature point detection algorithm[J].Computer Engeering,2011,37(13):196-198(in Chinese).[张永,纪东升.一种改进的Harris特征点检测算法[J].计算机工程,2011,37(13):196-198.]
[7]HANG Long,GUO Li,LI Yuyun.SIFT algorithm parallel implementation and application[J].Computer Engineering and Applications,2010,46(20):56-59(in Chinese).[韩龙,郭立,李玉云.SIFT算法的并行实现及应用[J].计算机工程与应用,2010,46(20):56-59.]
[8]SHOU Zhaoyu,OUYANG Ning,ZHANG Huajun,et al.Research on real-time video stitching based on SURF and dynamic ROI[J].Computer Engineering and Design,2013,34(3):998-1003(in Chinese).[首照宇,欧阳宁,张华俊,等.基于SURF和动态ROI的实时视频拼接[J].计算机工程与设计,2013,34(3):998-1003.]
[9]Rosten E,Porter 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.
[10]Calonder M,Leptit V,Strecha C.BRIEF:Binary robust independent elementary features[G].LNCS 6314:Computer Vision-ECCV.Berlin:Springer Berlin Heidelberg,2010:778-792.
[11]Mair E,Hager G,Burschka D,et al.Adaptive and generic corner detection based on the accelerated segment test[G].LNCS 6312:Computer Vision-ECCV.Berlin:Springer Berlin Heidelberg,2010:183-196.
[12]Rublee E,Rabaud V,Konolige K,et al.ORB:An efficient alternative to SIFT or SURF[C]//In International Conference of Computer Vision,2011:2564-2571.
[2]HU Qiong,QIN Lei,HUANG Qingming.A survey on visual human action recognition[J].Chinese Journal of Computers,2013,36(12):2512-2524(in Chinese).[胡琼,秦磊,黄庆明.基于视觉的人体动作识别综述[J].计算机学报,2013,36(12):2512-2524.]
[3]FAN bo,YANG Xiaomei,HU Xueshu.Super-resolution image reconstruction algorithms based on compressive sensing[J].Journal of Computer Applications,2013,33(2):480-483(in Chinese).[樊博,杨晓梅,胡学姝.基于压缩感知的超分辨率图像重建[J].计算机应用,2013,33(2):480-483.]
[4]ZHANG Mingjie,KANG Baosheng.Modified object detection and automatic tracking method[J].Computer Engineering and Design,2014,35(4):1308-1311(in Chinese).[张明杰,康宝生.改进的目标检测与自动跟踪方法研究[J].计算机工程与设计,2014,35(4):1308-1311.]
[5]JIAO Lilong,HAN Xie,LI Dingzhu.Improved image mosaic algorithm based on feature points matching[J].Computer Engineering and Design,2014,35(3):918-922(in Chinese).[焦丽龙,韩燮,李定主.改进的基于特征点匹配的图像拼接融合算法[J].计算机工程与设计,2014,35(3):918-922.]
[6]ZHANG Yong,JI Dongsheng.Improved Harris feature point detection algorithm[J].Computer Engeering,2011,37(13):196-198(in Chinese).[张永,纪东升.一种改进的Harris特征点检测算法[J].计算机工程,2011,37(13):196-198.]
[7]HANG Long,GUO Li,LI Yuyun.SIFT algorithm parallel implementation and application[J].Computer Engineering and Applications,2010,46(20):56-59(in Chinese).[韩龙,郭立,李玉云.SIFT算法的并行实现及应用[J].计算机工程与应用,2010,46(20):56-59.]
[8]SHOU Zhaoyu,OUYANG Ning,ZHANG Huajun,et al.Research on real-time video stitching based on SURF and dynamic ROI[J].Computer Engineering and Design,2013,34(3):998-1003(in Chinese).[首照宇,欧阳宁,张华俊,等.基于SURF和动态ROI的实时视频拼接[J].计算机工程与设计,2013,34(3):998-1003.]
[9]Rosten E,Porter 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.
[10]Calonder M,Leptit V,Strecha C.BRIEF:Binary robust independent elementary features[G].LNCS 6314:Computer Vision-ECCV.Berlin:Springer Berlin Heidelberg,2010:778-792.
[11]Mair E,Hager G,Burschka D,et al.Adaptive and generic corner detection based on the accelerated segment test[G].LNCS 6312:Computer Vision-ECCV.Berlin:Springer Berlin Heidelberg,2010:183-196.
[12]Rublee E,Rabaud V,Konolige K,et al.ORB:An efficient alternative to SIFT or SURF[C]//In International Conference of Computer Vision,2011:2564-2571.