计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (23): 173-178.

• 图形图像处理 • 上一篇    下一篇

基于FPGA的轮廓提取并行计算系统研究及实现

邵兴龙,柴志雷,张圆蒲,钟传杰   

  1. 江南大学 物联网工程学院,江苏 无锡 214122
  • 出版日期:2015-12-01 发布日期:2015-12-14

Study and implementation of FPGA-based parallel contour extraction system

SHAO Xinglong, CHAI Zhilei, ZHANG Yuanpu, ZHONG Chuanjie   

  1. School of IoT Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • Online:2015-12-01 Published:2015-12-14

摘要: 针对高质量的轮廓提取算法计算量大、实时性差的问题,提出了一种基于现场可编程门阵列(FPGA)的图像轮廓并行计算系统。通过设计适合的硬件结构及相应的算法改进,采用了多种不同的并行方式加速算法的计算。实现了一种高质量的轮廓提取算法——Pb(Probability Boundary)算法的高速计算。实验结果表明,在FPGA工作频率200 MHz时,被处理图像分辨率为481×321时,该系统处理速度可达39帧/s,为将Pb算法应用于实际系统提供了条件。

关键词: 现场可编程门阵列(FPGA), 图像轮廓提取, 概率轮廓(Pb)算法, 并行计算

Abstract: This paper proposes and implements a FPGA-based parallel image contour extraction system in order to deal with weak real-time performance of high-quality contour extraction algorithm processing. The proposed system is designed by adopting suitable hardware structures and necessary algorithm modifications, which is able to accelerate the contour extraction algorithm by providing different types of parallelisms. One high speed processing system for a state-of-the-art high-quality contour extraction algorithm, Pb (Probability Boundary) algorithm, is implemented in this paper. Experimental results show that this system is able to process 481×321 images at the frame rate 39 f/s, when the working frequency is 200 MHz. It provides a solid foundation to deploy Pb in real-world applications.

Key words: Field Programmable Gate Array(FPGA), image contour extraction, probability boundary algorithm, parallel computation