Vivado HLS嵌入式实时图像处理系统的构建与实现  

Construction and implementation of Vivado HLS embedded real-time image processing system

作　　者：张艳辉[1] 郭洺宇 何宾[1] 

机构地区：[1]北京化工大学信息科学与技术学院,北京100029 [2]北京卫星环境工程研究所,北京100094

出　　处：《电子技术应用》2016年第9期115-117,121,共4页Application of Electronic Technique

摘　　要：传统的基于CPU、GPU和DSP的处理平台难以满足图像实时处理的要求,而FPGA在并行图像处理上有着独一无二的优势,在性能和成本之间提供更加灵活的选择。通过Xilinx最新的Vivado HLS工具,设计实现了可变参数的拉普拉斯算子图像滤波算法,并且在ZYNQ-7000 So C上构建了可视化的实时嵌入式图像处理系统。实验结果表明,系统可以实现不同的图像处理算法,很好地满足了图像处理的实时性、高性能、低成本要求,对未来高性能图像处理系统的设计和实现提供了很好的借鉴。Traditional processing platforms which are based on CPU, GPU and DSP cannot meet the requirements in real- time processing. FPGA has unique advantages in parallel image processing and provides a much more flexible option between performance and power consumption. By using Vivado High Level Synthesis（ HLS）, tool of Xilinx, a variable parameter laplacian image filter algorithm has been implemented. And a kind of real- time and visual embedded image processing system is constructed on the ZYNQ- 7000 So C. The experimental results of this system verify different image processing algorithm can be implemented easily on this platform, requirements of real- time, high performance and low cost can be satisfied well. This design provides reference for further high performance image processing system.

关 键 词：VIVADO HLS 图像处理 拉普拉斯算子 ZYNQ-7000 嵌入式Linux QT 

分 类 号：TP368.2[自动化与计算机技术—计算机系统结构]