Dense 3D map building based on LRF data and color image fusion

Author

Ohno, Kazunori ; Tadokoro, Satoshi

Author_Institution

Graduate Sch. of Inf. Sci., Tohoku Univ., Sendai, Japan

fYear

2005

fDate

2-6 Aug. 2005

Firstpage

2792

Lastpage

2797

Abstract

Research objective of the authors is 3D map building and localization of search robot for rescue use. In this paper, the authors propose a novel method of dense 3D map building and present its trial result. For building a map, it is necessary to estimate robot motion. However, on rubble, it is difficult to estimate robot motion by using odometry or gyro. Therefore, in this framework, rough 3D map and discrete robot motions are derived using SLAM based on 3D scan matching. ICP algorithm is used for the matching method. Then, the dense 3D map is reconstructed from the rough 3D map and texture images.

Keywords

emergency services; image colour analysis; image matching; image reconstruction; image texture; mobile robots; motion estimation; robot vision; stereo image processing; telerobotics; 3D scan matching; ICP algorithm; LRF data; SLAM based; color image fusion; dense 3D map building; gyroscope; map reconstruction; odometry; rescue robot; robot localization; robot motion estimation; search robot; space carving; texture images; Color; Earthquakes; Image reconstruction; Iterative closest point algorithm; Motion estimation; Orbital robotics; Robot motion; Seismic measurements; Shape measurement; Simultaneous localization and mapping; Dense 3D Map; ICP; LRF; SLAM on rubble; Space Carving;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Robots and Systems, 2005. (IROS 2005). 2005 IEEE/RSJ International Conference on

Print_ISBN

0-7803-8912-3

Type

conf

DOI

10.1109/IROS.2005.1545235

Filename

1545235