This document describes a rear vehicle detection system using computer vision techniques for collision avoidance. The system aims to detect vehicles from a mounted camera on a host vehicle and alert the driver of an impending rear-end collision. It utilizes the fast radial symmetry transform to detect the blooming effect of vehicle rear lights under various illumination conditions. Candidate light pairs are identified and their symmetry and distance from the camera are estimated. The system achieves high detection rates on benchmark datasets, including in adverse weather. Future work involves correlating vehicle and braking detection with estimating driver attention level.