摘要
提出了一种基于双边滤波的图像边缘检测方法。利用图像像素的空间邻近度和灰度相似度的乘积来代替传统的Canny算法中高斯滤波的权系数,用该乘积与原图像进行卷积运算,然后通过非极大值抑制和高低阈值的方法检测出图像的边缘。基于双边滤波的图像边缘检测方法不仅有效地去除了图像中的噪声,而且很好地检测出了图像的真实边缘。
In this paper, a new method of edge detection based on bilateral filtering is presented. This method makes use of the product of geometric closeness and photometric similarity replace coefficient of Gaussian filter in Canny arithmetic. This product convolves with image and uses non- maxima suppression,two threshold methods detect boundary of image. Edge detection based on bilateral filtering method not only filters noise,but also effectively detects the real boundary of image.
出处
《计算机技术与发展》
2007年第4期161-163,共3页
Computer Technology and Development
基金
空军预研基金资助项目(HTEHY2002013)
关键词
双边滤波
图像平滑
边缘检测
非极大值抑制
bilateral filtering
image smcothing
edge detection
non - maxima suppression
