Object recognition using sparse, reduced-dimension point cloud data

Abstract

It has been shown that point separations as a feature derived from point clouds can be used to discriminate between two objects of similar class. Here we show that the same feature derived from sparse point clouds can maintain significant discrimination capability. Using the point-separation feature, templates created from random realizations of a point cloud are developed for several vehicles. The templates are then used in two-class discrimination tests. The point-separation feature is shown to produce reliable discrimination using a log-likelihood ratio between two objects.

A fast and feature-retained simplification method for point cloud

“…Point cloud is a recording form of object's surface information which is obtained by LiDAR. 3D point cloud has been widely used in urban planning, agricultural development, water conservancy engineering, unmanned driving, industrial measurement, industrial modeling and other aspects due to its characteristics of collecting conveniently, high precision and noncontact measurement [1][2][3][4][5][6]. The number of points in the raw point cloud collected by LiDAR scanner is huge, and the huge number of points greatly increases the difficulty of data processing.…”