An information integration system of signal intersection with multi-mode traffic data collection and analysis

A traffic information integration system of signal intersection with a multi-mode data collection joint-method is presented. Data acquisition modes at signal intersections contain the one with vehicle loop detector and traffic signal controller, and the other based on real-time video technique. Five basic parameters can be estimated by analyzing flow data from loop detectors with traffic signal timing. A video micro-detection for detecting directional vehicle flow in a crossing is improved to overcome issues as fake background, vehicle sheltering etc., with the combination of active search, passive search and block-matching with modified predictive line search (PLS). In addition, a macro-detection with enhanced morphological edge processing and scalable-window algorithm is proposed to detect the vehicle queue of approaches. Finally, the established intelligent platform of implemented information integration system provides a comprehensive traffic data dictionary, supporting for design, management and planning of signal intersections.