A Novel Recognition Algorithm in 3D Point Clouds based for on Local Spherical Harmonics

This paper presents a novel recognition algorithm of a 3D object in point clouds based on Local Spherical harmonics. In the proposed algorithm, the 3D point cloud of an object is decomposed into a set of local fields which constitute an orthogonal basis of expansion coefficients by Spherical Harmonic Expansion. The similarity between any corresponding local fields from two objects is expressed by a Euclidean distance between their expansion coefficients. The proposed algorithm aims to, provide a method to solve the problem of incomplete point cloud recognition. Our algorithm outperforms the existing approaches including Iterative Closest Point (ICP) and Discriminant Shape Primitives (DSP) with a recognition rate of 95.1% on the extension of Princeton Shape Benchmark and it has achieved a recognition rate of 92.9% on the extension of UWA Data-set.