Hardware Optimization for Real-Time Target Tracking using Fast-Normalized Cross Correlation
Omer Khan, Nayab Saeed, Raheel Muzzammel, Umair Tahir, Omar Azeem · 2019
Target tracking in real time is existing challenge for developers. The algorithms for target tracking require some programmable embedded platform for implementation. In this research “TMS320DM642 evaluation module with TVP video decoders” digital signal processor (DSP) embedded board is chosen as platform for implementation of target tracking algorithm. Optimization of real time target tracking is limited to some extend for all techniques. These techniques with software constraints can be improved with hardware optimization algorithm. Hence major focus of this research is to monitor results of hardware optimization algorithm. These optimization techniques provides efficient utilization of existing hardware resources. Various techniques like central processing unit (CPU) optimization, memory optimization and minimization of code are applied on target tracking algorithm using fast normalized cross correlation. Results are then investigated and studied to observe the enhancement in Process time, frame rate, accuracy and power consumption.