Multiple range image registration by matching local log-polar range images

Authors:
Takeshi Masuda
Affiliations:
National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan
Venue:
ACCV'06 Proceedings of the 7th Asian conference on Computer Vision - Volume Part I
Year:
2006

Citing 11
Cited 0

Rigid body motion from range image sequences

CVGIP: Image Understanding
Structural Indexing: Efficient 3-D Object Recognition

IEEE Transactions on Pattern Analysis and Machine Intelligence - Special issue on interpretation of 3-D scenes—part II
A Method for Registration of 3-D Shapes

IEEE Transactions on Pattern Analysis and Machine Intelligence - Special issue on interpretation of 3-D scenes—part II
Rigid, affine and locally affine registration of free-form surfaces

International Journal of Computer Vision
Point Signatures: A New Representation for 3D Object Recognition

International Journal of Computer Vision
Using Spin Images for Efficient Object Recognition in Cluttered 3D Scenes

IEEE Transactions on Pattern Analysis and Machine Intelligence
Multidimensional binary search trees used for associative searching

Communications of the ACM
Direct Estimation of Range Flow on Deformable Shape From a Video Rate Range Camera

IEEE Transactions on Pattern Analysis and Machine Intelligence
Automatic crude patch registration: toward automatic 3D model building

Computer Vision and Image Understanding - Registration and fusion of range images
Differential invariants as the base of triangulated surface registration

Computer Vision and Image Understanding - Registration and fusion of range images
Registration and integration of multiple range images by matching signed distance fields for object shape modeling

Computer Vision and Image Understanding - Registration and fusion of range images

Quantified Score

Hi-index	0.00

Visualization

Abstract

We propose a method for coarse registration of multiple range images. A local log-polar range image is computed at every surface point of all input range images, and an invariant feature vector is generated from it. The correspondence of point pairs is determined by finding the closest feature vector pairs derived from different range images. The correspondence is validated, and the RANSAC is applied for extracting inlier point pairs to determine pairwise transformations between input range images. Finally, the global registration is determined by construcing the view tree of the input range images. The result of coarse registration is used as the initial state for the fine registration which is followed by the object shape modelling.