Drivers who run red lights are a menace, causing about 260,000 accidents and taking about 750 lives every year in the United States. One good way to damp down on redlight runners is installing video cameras that capture violators on film and record their license plate numbers, so that tickets can be issued. But those cameras are expensive, costing at least $50,000. So when a municipality wants to use them, it helps to know which intersections need the most monitoring.
Enter University of Florida civil engineering professors Scott Washburn and Ken Courage. Along with graduate student Shaun MacKenzie, they're working on a prototype system that combines video and audio equipment with computer software to automatically, accurately, and inexpensively identify those intersections with the highest number of incidents.
The system uses video cameras and audio encoders that can pick up the eight different tones that a traffic signal can emit at a four-way intersection. It also uses detectors at the entrance, exit, and midway point of the crossroads to determine when cars have gone through it after the light is red. The result is an accurate count of the number of cars running red lights.