A Roadside Camera-Radar Sensing Fusion System for Intelligent Transportation

Cooperative Vehicle Infrastructure System (CVIS) is one of the most valued technologies in intelligent transportation system. By use of roadside sensing, it provides extended coverage and more dimensions traffic information compared with independent perception by vehicle sensors. Sensing fusion of camera and radar can overcome the shortcomings of both sensors so that it is the trend of roadside sensing in CVIS. In this paper, we propose a novel roadside sensing fusion system. It filters out background objects from radar detection to avoid wrong calibration and fusion with camera detection, and makes calibration of camera and radar automatically to reduce time cost of system implementation. Experiment results show that it can be fast implemented to automatically acquire accurate roadside sensing information.