Traditional shape from shading techniques, using a single image, do no
t reconstruct accurate surfaces and have difficulty with shadow areas.
Traditional shape from photometric stereo techniques have the disadva
ntage that they need all of the input images together at once to minim
ize the total cost, and this process must be restarted if new images b
ecome available. To overcome the shortcomings of the above two techniq
ues, we introduce a new technique called shape from photomotion. Shape
from photomotion uses a series of 2-D Lambertian input images, genera
ted by moving a light source around a scene, to recover the depth map.
In each of the input images, the object in the scene remains at a fix
ed position and the only variable is the light source direction. The m
ovement of the Light source causes a change in the intensity of any gi
ven point in the image. The change in intensity is what enables us to
recover the unknown parameter, the depth map, since it remains constan
t in each of the input images. This configuration is suitable for iter
ative refinement through the use of the extended Kalman filter. Our no
vel method for computing shape is a continuous form of the photometric
stereo technique. It significantly differs from photometric stereo in
the sense that the shape estimate will not only be computed for each
light source orientation, but also gradually be refined by photomotion
. Since the camera is fixed, the mapping between the depths at various
light source locations is known; therefore, this method has an advant
age over those which move the camera (egomotion) and keep the light so
urce fixed. Results of this method are presented for sequences of synt
hetic and real images. (C) 1996 Academic Press, Inc.