Edge Detection in Optical Flow Fields

Book One

Volume

Proceedings of the AAAI Conference on Artificial Intelligence, 2

Vision

Optical flow is potentially valuable as a source of spatial information. Current techniques provide flow fields which are noisy and sparse, making the recovery of spatial proper ties difficult at best. This paper describes a technique for locating discontinuities in optical flow, which typically correspond to object boundaries. A simple blurring interpolator is used to smooth out noise and produce denser fields. Discontinuities are found by locating the vector field equivalent of zero crossings in the Laplacian of a scalar field. The technique is illustrated by applying it to realistic vector fields which are both noisy and sparse.