Traffic flow detection based on scene knowledge

摘要

Traffic flow detection is the advanced stage in the intelligence surveillance. In this paper, we first parameterize the scene knowledge (as lane etc.) using Radon transform and Polynomial fitting to boundary of road. The features of Hu moment were used for tracking the moving objects; the scene knowledge and the fictitious coil were cited for analyzing the traffic flow of moving objects. The experimental results show that our algorithms are effectives.