摘要
针对目前ICP算法存在因初始变换矩阵的不确定性造成收敛速度较慢、点云间对应点难以精准匹配造成计算复杂度高的问题,提出一种利用点云局部坐标系,以关键点的邻域内点集的旋转体积为特征对关键点进行表述的描述子。藉此实现点云间关键点点对的准确匹配,达到粗配准的目的,提高点云配准的效率。分析实验结果表明,该描述子相比于其它描述子具有速度快、旋转平移鲁棒性高、抗噪性好的特点,能够应用于实际点云配准和对象识别。
For the ICP algorithm,the convergence speed is slow due to the uncertainty of the initial transformation matrix,and the correspondence points in the point cloud are difficult to match accurately,leading to high computational complexity.Therefore,a kind of descriptor based on local reference frame and the rotational volume of their neighborhood was put forward.The accurate matching of pairwise key points between point clouds was obtained using the proposed feature descriptor.The purpose of coarse registration was achieved,and the correspondence key points improved the efficiency of point cloud registration.Experimental results show that the descriptor has the characteristics of high speed,high robustness and good noise immunity,which can be applied to point cloud registration and object recognition.
引文
[1]Chen J,Belaton B.An improved iterative closest point algorithm for rigid point registration[J].Communications in Computer&Information Science,2014,481:255-263.
[2]Tagliasacchi A,Schr9der M,Tkach A,et al.Robust articulated-ICP for real-time hand tracking[J].Computer Graphics Forum,2015,34(5):101-114.
[3]GUO Yulan,LU Min,TAN Zhiguo,et al.Survey of local feature extraction on range images[J].Pattern Recognition and Artificial Intelligence,2012,25(5):783-791(in Chinese).[郭裕兰,鲁敏,谭志国,等.距离图像局部特征提取方法综述[J].模式识别与人工智能,2012,25(5):783-791.]
[4]Guo Y,Sohel F,Bennamoun M,et al.Rotational projection statistics for 3dlocal surface description and object recognition[J].International Journal of Computer Vision,2013,105(1):63-86.
[5]Guo Y,Bennamoun M,Sohel FA,et al.3Dfree form object recognition using rotational projection statistics[C]//Applications of Computer Vision.FL:IEEE,2013:1-8.
[6]Guo Y,Sohel F A,Bennamoun M,et al.RoPS:A local feature descriptor for 3Drigid objects based on rotational projection statistics[C]//International Conference on Communications,Signal Processing,and Their Applications.Sharjah:IEEE,2013:1-6.
[7]Rusu RB,Bradski G,Thibaux R,et al.Fast 3Drecognition and pose using the Viewpoint Feature Histogram[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems.Taipei:IEEE,2014:2155-2162.
[8]Li P,Wang J,Zhao Y,et al.Improved algorithm for point cloud registration based on fast point feature histograms[J].Journal of Applied Remote Sensing,2016,10(4):045024.
[9]Salti S,Tombari F,Di stefano L.Shot:Unique signatures of histograms for surface and texture description[J].Computer Vision and Image Understanding,2014,125(8):251-264.
[10]Zaharescu A,Boyer E,Horaud R.Keypoints and local descriptors of scalar functions on 2dmanifolds[J].International Journal of Computer Vision,2012,100(1):78-98.
[11]Petrelli A,Stefano LD.On the repeatability of the local reference frame for partial shape matching[C]//IEEE International Conference on Computer Vision.Barcelona:ICCV,2011:2244-2251.