基于角不变特征的三维点云配准算法RegistrationforDPointCloud.PDF

基于角不变特征的三维点云配准算法 基于角不变特征的三维点云配准算法 Registration for 3-D Point Cloud Using Angular-Invariant Feature 姜军  程俊  陈兴林 摘 要  本文针对三维配准中的对应点选择问题，提出了一种角不变特征。特征量是一个 维向量，向量中的元 k 素表示点的法向量与该点的 个最近邻的法向量之间的夹角。这种角特征不随尺度、旋转变换而变化，能够应 k 用到表面曲率较小的配准问题。角特征不需要给定初始变换矩阵和平移向量，因而提高了配准的收敛速度，降 低了配准误差。此外，角不变特征不要求严格的采样策略。本文通过证明了基于角不变特征的配准算法优ICP 算法，以及基于曲率不变特征的算法。 关键词  三维配准；ICP算法；角不变量；曲率不变量；三维点云 ABSTRACT This paper proposes an angular-invariant feature for 3-D registration procedure to perform reliable selection of point correspondence. The feature is a k-dimensional vector, and each element within the vector is an angle between the normal vector and one of its k nearest neighbors. The angular feature is invariant to scale and rotation transformation, and is applicable for the surface with small curvature. The feature improves the convergence and error without any assumptions about the initial transformation. Besides, no strict sampling strategy is required. Experiments illustrate that the proposed angular-based algorithm is more effective than iterative closest point (ICP) and the curvature-based algorithm. KEYWORDS 3-D registration; ICP; angular invariant; curvature invariant; 3-D point cloud 1 Introduction hape inspection of objects from real world typically search for correspondences. Soucy et al. [7] locally Srequires three stages:(1)the data capture stage to register surface patches by minimizing the distance sample the 3-D world using a range camera;(2)the data between Darboux frames over an entire neighborhood. registration stage to align the sampled data shape to the Yang et al. [8] minimize a scaled product of positional and model shap