A 3D classifier trained without field samples

Abstract

This paper presents a 3D classifier that is shown to maintain performance whether trained with real sensor data from the field or purely trained with 3D geometric (Computer Aided Design, CAD, like) models (downloaded from the Internet for instance). The proposed classifier is a global 3D template matching technique which exploits the location of the ground surface for more accurate alignment. The segmentation and position of the ground is given by the segmentation technique in [7] (which does not assumed the ground to be flat). The proposed classifier outperforms Spin Image and Fast Point Feature Histogram (FPFH) based classifiers by up to 30% (the latter being tested at different scales), in the case of sparse 3D data acquired with a Velodyne sensor. In addition, the experimental results suggest that field samples may not be required in the training set of alignment-based 3D classifiers. This finding implies that the laborious task of gathering hand labelled field data for training may be avoidable for this type of classifier.

Dec-MCTS: Decentralized planning for multi-robot active perception

“…The observation likelihood is calculated by measuring the similarity between the shape of the point cloud with each model instance in the database. Similarity is computed by first aligning the point clouds of a pair of objects using the iterative closest point (ICP) algorithm (Besl and McKay, 1992) and then calculating the symmetric residual error (Douillard et al, 2012). Objects may merge or split after each observation if the segmentation changes.…”

Decentralised Monte Carlo Tree Search for Active Perception

“…The posterior probability distribution for each object after a set of observations is computed using recursive Bayes' rule. The observation likelihood is computed as the symmetric residual error of an iterative closest point (ICP) alignment [8] with each model in the database. Objects may merge or split after each observation if the segmentation changes.…”