A real-time computer vision system for measuring traffic parameters

摘要: For the problem of tracking vehicles on freeways using machine vision, existing systems work well in free-flowing traffic. Traffic engineers, however, are more interested monitoring when there is congestion, and current break down for congested traffic due to partial occlusion. We developing a feature-based approach task under congestion. Instead entire vehicles, vehicle sub-features tracked make system robust In order group together that come from same vehicle, constraint common motion used. this paper we describe system, real-time implementation network DSP chips, experiments approximately 44 lane hours video data.