M. Otte et Hh. Nagel, ESTIMATION OF OPTICAL-FLOW BASED ON HIGHER-ORDER SPATIOTEMPORAL DERIVATIVES IN INTERLACED AND NONINTERLACED IMAGE SEQUENCES, Artificial intelligence, 78(1-2), 1995, pp. 5-43
Citations number
56
Categorie Soggetti
Computer Sciences, Special Topics","Computer Science Artificial Intelligence",Ergonomics
This contribution(1) investigates local differential techniques for es
timating optical flow and its derivatives based on the brightness chan
ge constraint. By using the tensor calculus representation we build th
e Taylor expansion of the gray-value derivatives as well as of the opt
ical flow in a spatiotemporal neighborhood. Such a formulation provide
s a unifying framework for all existing local differential approaches
and allows to derive new systems of equations for the estimation of th
e optical flow and of its derivatives. We also tested various optical
flow estimation approaches on real image sequences recorded by a calib
rated camera which was fixed on the arm of a robot. By moving the arm
of the robot along a precisely defined trajectory, we can determine th
e true displacement rate of scene surface elements projected into the
image plane and compare it quantitatively with the results of differen
t optical flow estimators. Since the optical flow estimators are based
on gray-value derivatives of up to fourth-order, we were forced to de
velop modified Gaussian derivative filters to obtain acceptable estima
tes for the derivatives, Further, we show quantitatively that these fi
lters contribute to a much more robust optical how estimation, In addi
tion, successive lines of TV-cameras have an offset in time due to the
interlace technique. We demonstrate the adaptation of filter kernels
for estimating higher-order spatiotemporal derivatives in interlaced i
mage sequences.