OPTIMIZED ARCHITECTURE DESIGN AND IMPLEMENTATION OF OBJECT TRACKING ALGORITHM ON FPGA

FPGA based Object tracking implementation is one of the most recent video
surveillance applications in embedded systems. In general, FPGA implementation is
more efficient than general purpose computers in attaining high throughput due to its
parallelism and execution speed. The system need to be designed on a standard frame
rate in such a way to achieve optimal performance in real time environment. Optimal
design of a system is dependent on minimizing the cost, area (device utility) and
power while achieving the required speed. Past research work that investigated object
tracking systems' implementation on FPGA achieved a significantly high throughput
but have shown high device utilization. This research work aims at optimizing the
device utilization under real time constraints. The Adaptive Hybrid Difference
algorithm (AHD), which is used to detect the moving objects, was chosen to be
implemented on FPGA due to its computation ability and efficiency with regard to
hardware implementation. AHD can work at various lighting conditions automatically
by determining the adaptive threshold in every period of time.