基于HVS的彩色图像边缘检测算子  

A Color Edge Detection Operator Based on Human Vision System

作　　者：林生佑[1] 石教英[1] 

机构地区：[1]浙江大学CAD&CG国家重点实验室,杭州310027

出　　处：《中国图象图形学报（A辑）》2005年第1期43-47,F008,共6页Journal of Image and Graphics

基　　金：国家自然科学基金重点项目(60033010);国家创新群体科学基金项目(60021201)

摘　　要：边缘检测是数字图像处理中的基本预处理方法之一,由于利用检测出来的图像边缘可以对图像做进一步的处理,因此它是图像处理中的基础算法。大家知道,边缘检测问题,其本质上是颜色差分的计算问题。尽管边缘检测在灰度图像处理当中得到了深入的研究,但对于彩色图像仍然是一个难题。为了能够较好地对彩色图像进行边缘检测,结合人类视觉系统(humanvisionsystem, HVS)的特点,通过分离颜色的亮度信息和色度信息,并通过压制次要信息来强调其中的重要信息,提出了一种新的彩色图像边缘检测算法。该算法先分别计算两个颜色的亮度距离和色度距离,然后将这两个距离的加权平均值作为最终的颜色距离。由于它能较好地检测出图像边缘,因此是一种实用有效的彩色图像边缘检测方法。Edge detection is one of the basic pre processing methods in digital image processing and computer vision. It is one of the most important algorithms making the processing going further. Edge detection is essentially a problem of color difference computing. Although edge detection has been explored deeply in gray image, it is still a difficult problem in color image for its trouble to define the color difference including not only value, but also direction. In this paper, a new color edge detection algorithm is proposed. This algorithm is based on Human Vision System (HVS) separating the chroma and intensity information of RGB color to emphasize the more important one. Inspired by this important property of HVS, the intensity difference and chroma difference are computed respectively, and then the weighted mean over these two differences is taken as the final color difference. This algorithm solves some problems produced by the existing algorithms and is proved to be a practical and effective color edge detection algorithm.

关 键 词：彩色图像边缘检测 HVS 灰度图像 测算法 数字图像处理 人类视觉系统 算子 亮度 颜色 色度 

分 类 号：TB851.7[一般工业技术—摄影技术] TP391.41[自动化与计算机技术—计算机应用技术]