FPGA Based Real-Time Target Tracking on a Mobile Platform

Abstract

An embedded video processing system is explained which tracks an object in real-time through a parallel processing architecture designed for an FPGA. Mechanical assembly of the particular mobile robot we utilized is also discussed, although the system can easily be ported to any mobile platform. The video processing is completely based on point function, which allows for straight-forward parallelization. The video processing system can be used to follow any selected target on the basis of its chrominance properties (Chrome Blue-Cb and Chrome Red-Cr). The adaptability of the robot for different lighting conditions is another key contribution.