Using multisensory images to derive the structure of three-dimensional objects - A review

Deriving object descriptions which represent 3-dimensional structures has been approached using multiple 2-dimensional intensity images of a scene and, more recently, by employing specialized range sensing hardware that can directly measure all the spatial coordinates of points in a scene. This paper presents a brief synopsis of efforts using only intensity images and a more extensive review of research that incorporates range imagery. The strengths and weaknesses of both intensity and range domains are discussed with an eye toward combining them into a system that uses the advantages of each domain to offset the disadvantages of the other.