free formed surfaces. The method using meta-balls0' A new robust method for calculating curvature vet- can represent complex free formed surfaces by combitors from range data is proposed. A new schenle ISL (is0 nation of simple functions. It has problems in regard to slant loop) is introduced for this method. An ISL is de- the type of objects and uniqueness of its represenlafined for each target data point as each of its compo- tion. nents has a normal vector which has a constant angle Other methods decompose the surface into patches. against the normal vector of the target. Usually, the loop has an ellipse-like shape of which the primary and secondary axis match the curvature vectors. We applied our nlethod to a synthesized data (ellipsoids) with synthesized random noise. The calculated vectors are compared with theoretical curvature vectors. One of other traditional methods (quadric patch) is also tested to the same data and the results are reported. These experimental results show that our method is robust with noisy data.
[1]
R. R. Martin,et al.
Differential geometry applied to curve and surface design
,
1988
.
[2]
Ramesh C. Jain,et al.
Three-dimensional object recognition
,
1985,
CSUR.
[3]
James F. Blinn,et al.
A Generalization of Algebraic Surface Drawing
,
1982,
TOGS.
[4]
Ramesh C. Jain,et al.
Segmentation through Variable-Order Surface Fitting
,
1988,
IEEE Trans. Pattern Anal. Mach. Intell..
[5]
Nabih N. Abdelmalek,et al.
Algebraic error analysis for surface curvatures and segmentation of 3-D range images
,
1990,
Pattern Recognit..
[6]
Takashi Yoshimi,et al.
Multi Light Sources Range Finder System
,
1991,
MVA.
[7]
Ruzena Bajcsy,et al.
Recovery of Parametric Models from Range Images: The Case for Superquadrics with Global Deformations
,
1990,
IEEE Trans. Pattern Anal. Mach. Intell..