Single-Object Detection Hardware Accelerator Using XfOpenCV Library

Real-time image processing plays an important role in object detection in various fields, such as electric vehicle and security system. Image processing contains complex algorithms that require more processing time and consume lots of power for a CPU to solve. On the other hand, field programmable gate array (FPGA) having high computational power and capability of working with CPU solves the complex algorithm with more speed and less power consumption. Utilization of energy and resource is important parameter to make any system effective. In this paper, we develop an intellectual property (IP) for single-object detection using xfOpenCV library APIs. The detection contains libraries such as thresholding, Fast corner and Boundary Scan. The aim is to increase the ability of data processed by the system and save the power consumed by CPU by using FPGA. Such effective system can be used where power plays an important role and limited battery is available such as in electric vehicle. The targeted board is Zybo Zynq7000, which is more reliable and faster than CPU.