Fast Visual Feature Selection and Tracking in a Hybrid Reconfigurable Architecture

Abstract

We present a fast visual feature tracking system which takes advantage of dedicated hardware to perform the computationally intensive step of selection. A software system uses the output of the hardware selector to develop tracks using filtering and data association techniques, and image-based validation. By using separate, custom hardware to select hundreds of points per frame then associating these points to tracks in software, we achieve greater than 60Hz real-time tracking.