图像预处理算法的FPGA实现  

FPGA-based real-time image preprocessing system

作　　者：祁欣 陈剑锋 罗伟林[1] Qi Xin;Chen Jianfeng;Luo Weilin(College of Mechanical Engineering and Automation,Fuzhou University,Fuzhou 350108,China)

机构地区：[1]福州大学机械工程及自动化学院,福州350108

出　　处：《国外电子测量技术》2021年第2期102-107,共6页Foreign Electronic Measurement Technology

基　　金：福建省海洋高新产业发展专项(闽海洋高新[2016]16号)资助。

摘　　要：针对图像数据量大、噪点多、边缘轮廓模糊等关键技术问题,设计了一套基于现场可编程门阵列(FPGA)的实时图像预处理系统,并在FPGA上实现了图像预处理综合算法。采集到的图像经过灰度转换算法,降低了数据量;边缘提取前加入中值滤波算法,去除了噪点;边缘提取后,为了使边缘的连贯性和轮廓加强,加入了形态学滤波算法,保证后续图像处理的质量。对所设计的图像预处理系统在实际环境下进行测试,输出显示的效果不仅满足实时性和处理速度的要求,而且实现了图像去噪、数据量小、边缘细节完整的图像预处理功能。A FPGA-based real-time image preprocessing system is proposed to deal with technical issues on large amount of data,noise,and blurred edge contours of image.Gray transformation is conducted to decrease the data captured.Edge detection is performed to reflect the characteristic of the image.Before edge detection,median filtering is conducted to eliminate the noises in the image and morphological filtering is carried out to improve the quality of the image after edge detection.Experiment is performed to confirm the proposed image processing system.Results show that the real-time performance and processing rate can be guaranteed.The processing system is also characterized by little noise,small data size,and integrity of the edge details.

关 键 词：FPGA 中值滤波 边缘检测 形态学滤波 时序 实时性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]