Issue No. 04 - April (2009 vol. 31)
Kwang-Ho Bae , Curtin University of Technology, Perth
David Belton , Curtin University of Technology, Perth
Derek D. Lichti , University of Calgary, Calgary
We present a novel closed-form expression of positional uncertainty measured by a near-monostatic and time-of-flight laser range finder with consideration of its measurement uncertainties. An explicit form of the angular variance of the estimated surface normal vector is also derived. This expression is useful for the precise estimation of the surface normal vector and the outlier detection for finding correspondence in order to register multiple three-dimensional point clouds. Two practical algorithms using these expressions are presented: a method for finding optimal local neighbourhood size which minimizes the variance of the estimated normal vector and a resampling method of point clouds.
error analysis, Range data, Eigenvalues and eigenvectors
D. Belton, D. D. Lichti and K. Bae, "A Closed-Form Expression of the Positional Uncertainty for 3D Point Clouds," in IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 31, no. , pp. 577-590, 2008.