Accurate object localization in 3D laser range scans

Author

Nüchter, Andreas ; Lingemann, Kai ; Hertzberg, Joachim ; Surmann, Hartmut

Author_Institution

Inst. for Comput. Sci., Osnabruck Univ.

fYear

2005

fDate

18-20 July 2005

Firstpage

665

Lastpage

672

Abstract

This paper presents a novel method for object detection and classification in 3D laser range data that is acquired by an autonomous mobile robot. Unrestricted objects are learned using classification and regression trees (CARTs) and using an Ada Boost learning procedure. Off-screen rendered depth and reflectance images serve as an input for learning. The performance of the classification is improved by combining both sensor modalities, which are independent from external light. This enables highly accurate, fast and reliable 3D object localization with point matching. Competitive learning is used for evaluating the accuracy of the object localization

Keywords

image classification; laser ranging; learning (artificial intelligence); mobile robots; object detection; regression analysis; robot vision; stereo image processing; 3D laser range data; 3D laser range scan; 3D object localization; Ada Boost learning procedure; accurate object localization; autonomous mobile robot; classification tree; competitive learning; external light; object classification; object detection; point matching; reflectance image; regression tree; sensor modality; unrestricted object learning; Cognitive robotics; Computer science; Intelligent robots; Intelligent systems; Knowledge based systems; Laser theory; Mobile robots; Object detection; Regression tree analysis; Rendering (computer graphics);

fLanguage

English

Publisher

ieee

Conference_Titel

Advanced Robotics, 2005. ICAR '05. Proceedings., 12th International Conference on

Conference_Location

Seattle, WA

Print_ISBN

0-7803-9178-0

Type

conf

DOI

10.1109/ICAR.2005.1507480

Filename

1507480