Increasing driving safety with a multiple vehicle detection and tracking system using ongoing vehicle shadow information

This paper proposes a vehicle detection and tracking system based on processing monochrome images captured by a single camera. The work has mainly been focused on detecting and tracking vehicles in daylight conditions, viewed from inside a vehicle. Unlike previous work, this approach uses vehicle shadow clues and vehicle edge information to obtain cost effective and fast estimation. The proposed method includes road area finding which has been implemented by a lane detection algorithm to avoid false detections of vehicles caused by the distraction of background objects. Assuming that lanes are successfully detected, vehicle presence inside the road area is hypothesized by using "shadow" as a cue. Hypothesized vehicle locations are verified using vertical edges. After extracting vehicles, the algorithm effectively tracks them using a Kalman filter based tracking algorithm. A vehicle has been instrumented with various sensors for the experiments. Several sequences from real traffic situations have been tested, obtaining highly accurate multiple vehicle detections. Tracking information is used to estimate time-to-collision (TTC) and warn the driver for a possible collision.