Fast Simultaneous Alignment of Multiple Range Images Using Index Images

Authors:
Takeshi Oishi;Atsushi Nakazawa;Ryo Kurazume;Katsushi Ikeuchi
Affiliations:
University of Tokyo;Osaka University;Kyushu University;University of Tokyo
Venue:
3DIM '05 Proceedings of the Fifth International Conference on 3-D Digital Imaging and Modeling
Year:
2005

Citing 0
Cited 6

Flying Laser Range Sensor for Large-Scale Site-Modeling and Its Applications in Bayon Digital Archival Project

International Journal of Computer Vision
3D Model Segmentation and Representation with Implicit Polynomials

IEICE - Transactions on Information and Systems
Wavelet-Texture Method: Appearance Compression by Polarization, Parametric Reflection Model, and Daubechies Wavelet

International Journal of Computer Vision
Median Photometric Stereo as Applied to the Segonko Tumulus and Museum Objects

International Journal of Computer Vision
3D digital archive of the burghers of calais

VSMM'06 Proceedings of the 12th international conference on Interactive Technologies and Sociotechnical Systems
A coarse-to-fine IP-driven registration for pose estimation from single ultrasound image

Computer Vision and Image Understanding

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper describes a fast and easy-to-use simultaneous alignment method of multiple range images. The most time consuming part of alignment process is searching corresponding points. Although "Inverse calibration" method quickly searches corresponding points in complexity O(n), where n is the number of vertices, the method requires some look-up tables or precise sensorýs parameters. Then, we propose an easy-to-use method that uses "Index Image": "Index image" can be rapidly created using graphics hardware without precise sensorýs parameters. For fast computation of rigid transformation matrices of a large number of range images, we utilized linearized error function and applied incomplete Cholesky conjugate gradient (ICCG) method for solving linear equations. Some experimental results that aligned a large number of range images measured with laser range sensors show the effectiveness of our method.