, Assistant Professor, UMD-Electrical Engineering
Vehicle tracking processes on roads are computationally intensive. In the past, the different algorithms employed in vehicle tracking have been implemented using various software-based approaches. While software approaches have the advantage of flexibility in implementation and future modifications, the long computational time of these approaches often prevents real-time vehicle tracking from high-resolution spatial or temporal data. The goal of this project was to build a hardware-based vehicle tracking system that used a new algorithm based on vehicle motion detection and retained key elements of video-based tracking system design. It also used customized hardware whenever possible to shorten execution time, ultimately enabling real-time tracking at a high frame rate. The tracking system consisted of two major components: a hardware processor for vehicle motion detection, and a tracking algorithm based on motion estimation. Development of a hardware-based tracking system required three steps: 1) validation of the tracking algorithm using realistic video inputs; 2) implementing the algorithm in hardware; and 3) construction and testing. At the project's conclusion, tracking results had been largely validated and technical issues identified, such as sensitivity to camera jitter, which requires improvements to the algorithm to make it more robust. The design of custom hardware is currently underway, and testing will be carried out in the next phase of the project.
- Project number: 2008016
- Start date: 06/2007
- Project status: Completed
- Research area: Transportation Safety and Traffic Flow