Multi-sensor Measurements in Urban Environment

Within the project Sensor systems for urban environments, field trials with multi sensors in urban-like environments have been conducted. The sensor set includes video, IR, and acoustic sensors in complex scenarios with a number of actors. It is demonstrated how persons automatically can be detected, associated, and tracked in a video network with constrained communications and distributed algorithms. Acoustic sensors with high spatial resolution refines individual sound sources given the source location, which in turn can be used to associate the sound with persons or vehicles in video, for instance. A fusion model for IR and video is proposed, that in order to make surveillance systems more robust, co-utilize the richness of details and colors of the video with the invariance to shadows, reflections and darkness of the IR-camera. Also, trials with radar and IR for discovering concealed weapons have been conducted. Detection of concealed weapons using IR measurements has proved to be difficult. The conducted IR measurements show no traceable weapon contours. Radar, on the other hand, shows a significant difference comparing measurements with and without weapons, at least for somewhat large person-carried weapons like a bomb belt and an automatic rifle.