Argonauts:Iterative Closest Point (ICP) algorithm
From Wasteland
Contents |
Sun 30 October 2005
Spatio-temporal coordinates
WMVL, 11.00am to 1.00pm (exactly).
Attendees (in alphabetical order)
- Ramón Casero Cañas.
- Niranjan Joshi.
- Jeong-Gyoo Kim.
- Rohan Loveland.
Minutes
First meeting of the Argonauts ever. A legend is born. We have 120 min to learn all we can about the ICP as a crew. We find the classic paper:
P.J. Besl and N.D. McKay. A method for registration of 3-D shapes. IEEE Transactions on Pattern Analysis and Machine Intelligence, 14(2). Feb 1992.
We focus on the description of the algorithm in section IV.B. Quaternions are a cool way to represent rotations in 3D space. This paper is about rigid registration (rotation and translation) between two manifolds: one given as a sample of points, the other given as triangles, points, lines, etc.
The algorithm is relatively simple, but we just have no clue where (25) comes from, and it is important because the rotation-quaternion is computed from the matrix in (25) as the eigenvector with the largest eigenvalue.
It is decided that next time we will allow 15 extra minutes for conclusions.
Lunch in the Eagle and Child ensues.
--
