Real-time video processing in ZedBoard

• Main Author: Chua, Ivan Rui Xiang
• Other Authors: Arvind Easwaran
• Format: Final Year Project
• Language: English
• Published: 2017
• Subjects: DRNTU::Engineering::Computer science and engineering::Hardware::Input/output and data communications
• Online Access: http://hdl.handle.net/10356/70323
• Institution: Nanyang Technological University

Video processing is a technique that is widely used nowadays to meet the growing demands in the industry. By implementing a video processing system into a real-time operating system, it can be used in many ways such as mounting it on top of a drone to carry out search and rescue mission or for military usage, etc. During this operation, the video feed must be able to display the real-time scenario within a limited amount of time. Time is an essential and critical role in such operation. Thus, the system is required to complete all its operation within the given time frame. In this project, we focus on evaluating a hybrid processing platform such as Zynq 7000 (e.g. Avnet Zedboard) which comprise of both Commercial-Off-The-Shelf (COTS) processor (ARM) and Field Programmable Gate Array (FPGA) for a video processing real-time system. This report investigates the design of video processing system on FPGA and compares it against the implementation on a COTS processor. We integrate a low-cost Video Graphics Array (VGA) camera with Avnet Zedboard and implement a simple edge detection system on the FPGA. To evaluate the performance of our implementation, we implement the same system on Raspberry Pi (RPi) which comes with ARM Cortex-A7 processor. It is found that the performance of the video processing system implemented on FPGA is superior in comparison to that of COTS processor implementation. The comparison results show that FPGA has a much high frame per second (fps) than the ARM processor when processing video process operation.