Surveillance, Monitoring and Traffic Video Systems

Many methods have been proposed throughout the years
for solving multiframe video restoration problems.
With that, well-established restoration methods exist
only for situations in which all sources of blur and
degradation are known or easily predicted. When some
of the parameters are unknown the problem becomes
difficult. The presented research addresses this
challenge. At first motion estimation techniques are
evaluated. Based on the evaluation of these methods,
an improvement, through numerical exact derivation,
is suggested. Accurate evaluation of the motion field
allows a reliable video stabilization and improving
frames resolution. The super-resolution (SR) method
is then improved by the theory of discrete signal
reconstruction from sparse data. Additionally, SR s
potential and limitations are sought. Finally, the
developed stabilization and SR methods are utilized
for providing means for reasoning the scene observed.
The results presented are of both theoretical and
practical interest and offer new efficient tools for
substantial improvement of infrastructure of vision-
based systems in general and of intelligent
transportation systems in particular.

Autorentext

Barak Fishbain received his Ph.D in EE from Tel-Aviv University, Israel in Dec. 2008. His research interests are Computer Vision, Image Processing, Video Surveillance and Medical Imaging. Currently he is a postdoctoral fellow in the Dept. of Industrial Engineering and Operational Research in the University of California at Berkeley, USA

Klappentext

Many methods have been proposed throughout the years for solving multiframe video restoration problems. With that, well-established restoration methods exist only for situations in which all sources of blur and degradation are known or easily predicted. When some of the parameters are unknown the problem becomes difficult. The presented research addresses this challenge. At first motion estimation techniques are evaluated. Based on the evaluation of these methods, an improvement, through numerical exact derivation, is suggested. Accurate evaluation of the motion field allows a reliable video stabilization and improving frames resolution. The super-resolution (SR) method is then improved by the theory of discrete signal reconstruction from sparse data. Additionally, SR's potential and limitations are sought. Finally, the developed stabilization and SR methods are utilized for providing means for reasoning the scene observed. The results presented are of both theoretical and practical interest and offer new efficient tools for substantial improvement of infrastructure of vision-based systems in general and of intelligent transportation systems in particular.