FPGA based video processing system for ubiquitous applications

In this paper, a novel FPGA based video/image processing architecture is introduced that is designed to deal with advanced computer vision applications in ubiquitous systems. As basic components of this architecture, some low-level image feature extraction algorithms including the Sobel edge detector and SUSAN edge and corner detectors were designed and implemented using FPGA software tools. A simple PCFPGA combined platform was proposed and designed for testing the performance of our design. Experimental results showed these components worked very well on an FPGA chip.