A camera (32) captures successive images of light stripes (22) projected onto an object by a light curtain (18) positioned by a light source positioner (28). A background image is subtracted (616) therefrom, and the resulting image is boosted by binning (618), binarized with a thresholding algorithm (620), skeletonized (622), interpolated (624) and stored (626). Interpolated images are acquired for a plurality of light stripes (22). A processor (30) generates (1604) a 3-D surface model from Cartesian coordinates computed for non-zero camera pixels. A volumetric representation is determined (1610) from the offset of the object surface model relative to a model of a proximate surface, e.g. a seating surface (24). The object is classified (1614), e.g. by a trainable pattern recognitions system, responsive to 3-D shape descriptors (1606) of the 3-D surface model and to the volumetric representation (1610) or portions (1612) thereof. The detection of an occupant (14) in a vehicle (12) may be used to control a safety restraint system (36).