摘要
边缘检测是图像处理领域中的基本课题之一,广泛应用于图像分割和模式识别等领域.本文提出的基于模糊增强的Canny边缘检测方法是在模糊增强的基础上,用模糊增强代替原来Canny算法中的高斯滤噪,用最大类间方差法代替原来Canny算法中双阈值的设定和边缘连接.本文方法集中了以上几种算法的优点,从而获得了综合性能最优的边缘检测效果.利用本文介绍的方法作边缘检测,边界清晰,完整,所得图像在抗噪性能和保持信息完整性方面二者得以兼顾.通过调整参数可以满足不同光学效果的图片和不同处理目的的需要.
Edge detection is one of the basic subjects in image processing. It is widely used in image segmentation,mode identification,etc. This paper puts forward a novel Canny edge detection approach based on fuzzy enhancement. With this approach,Guass-filter, used in traditional Canny algorithm,is replaced by fuzzy enhancement, and double thresholds and edge connection are substituted by the maximum classification square error. The new approach involves in itself all the advantages of previous methods. It can make the detected edges much clearer and the antinoise performance of images much better. Moreover, the pictures with different optical Characteristics can be properly processed by adjusting some parameters so as to meet different requirements.
出处
《内蒙古工业大学学报（自然科学版）》
2008年第1期65-69,共5页
Journal of Inner Mongolia University of Technology：Natural Science Edition
