Model-Based 3D Point Cloud Segmentation for Automated Selective Broccoli Harvesting

Abstract

In this paper we address the topic of feature matching in 3D point cloud data for accurate object segmentation. We present a matching method based on local features that operates on 3D point clouds to separate crops of broccoli heads from their background. Our method outperforms recent methods based on 2D standard segmentation techniques as well as clustering spatial distances. We have implemented our approach and present experiments on datasets collected in cultivated broccoli fields, in which we analyse performance and capabilities of the system as a point feature-based segmentation method.