基于硬件描述语言的图像边缘特征提取技术研究  

Research on Image Edge Feature Extraction Technique Based on Hardware Description Language

作　　者：刘宗明[1,2,3] 张宇 曹姝清 卢山 叶东[1] 

机构地区：[1]哈尔滨工业大学电气工程及自动化学院,黑龙江哈尔滨150001 [2]上海航天控制技术研究所,上海200233 [3]上海市空间智能控制技术重点实验室,上海200233

出　　处：《红外》2015年第11期19-24,40,共7页Infrared

基　　金：国家自然科学基金面上项目(61374162);上海市科技人才计划(14XD1421400;14QB1401800)资助项目

摘　　要：对图像边缘特征的提取是图像处理中一项非常重要的研究内容。相比于角点信息,边缘信息对观测目标具有更丰富的表示含义,对目标外形轮廓的精确提取具有重要的研究意义。随着探测器分辨率的提升以及对数据更新率要求的提高,开发能快速高效地实现边缘提取的算法已成为人们研究的热点。着重研究了高精度图像处理算法的FPGA实现技术,并利用FPGA独有的并行流水线处理机制实现了高斯滤波、sobel算子和非极大值抑制等图像处理的基本运算。The extraction of image edge features is very important in the processing of images.Compared with the corner information,the edge information contains richer representation mean of the target to be observed.It is of significance to the accurate extraction of target contour.With the improvement of detector's resolution and the increase of data update rate,development of the algorithms for fast and effective edge extraction has been one of the research hotspots.The technique for implementing highprecision image processing algorithms using FPGA is studied emphatically.The basic image processing operations such as Gaussian filter,sobel operator and non-maxima suppression are implemented by using the unique parallel pipeline processing mechanism of FPGA.

关 键 词：硬件描述语言 图像处理 高斯滤波 SOBEL算子 非极大值抑制 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]