Algorithms for Image Analysis in Traffic Surveillance Systems

Abstract

The presence of various surveillance systems in many areas of the modern society is indisputable and the most perceptible are the video surveillance systems. This thesis mainly describes novel algorithm for vision-based estimation of the parking lot occupancy and the closely related topics of pre-processing of images captured under harsh conditions. The developed algorithms have their practical application in the parking guidance systems which are still more popular. One part of this work also tries to contribute to the specific area of computer graphics denoted as direct volume rendering (DVR).