Place Categorization for Outdoor Environment using Real-time Omni-directional Laser Scanner

Abstract

This paper presents a place categorization technique using a real-time omni-directional laser scanner for outdoor environment. Firstly, we created a large database consisting of range data of six categories, that is, coast, forest, indoor parking, outdoor parking, residential area, and urban area. Two kinds of features, spin-image and local binary pattern, are extracted from range data and used for the estimation of the category which the range data belongs to by SVM. The performance of the proposed technique is evaluated through the experiments using the developed database.