一种精确的自适应图像边缘提取方法  

An Exact Method for Adaptive Image Edge Detection

作　　者：黄剑玲[1] 邹辉[2] 

机构地区：[1]上饶师范学院数学与计算机系,江西上饶334000 [2]上饶职业技术学院机械系,江西上饶334000

出　　处：《计算机工程与科学》2009年第9期53-55,94,共4页Computer Engineering & Science

摘　　要：针对传统Canny算子需要人为指定参数和阈值等问题,提出一种自适应图像边缘提取方法。该方法使用改进的自适应中值滤波对图像去噪;用8点邻域梯度计算方法自适应计算滤波后图像梯度的幅值;根据图像的梯度信息特征,用梯度的熵和标准差自适应地调整高低阈值。实验结果表明,与传统Canny算子相比,该算法在滤除图像噪声、保留细节边缘和保持边缘连通性方面都获得了不错的效果,提高了边缘检测的自动化程度,更具有实用性。An adaptive edge detection method is proposed for the given parameters and thresholds of traditional Canny algorithms. This method eliminates the image noise by using an improved method of adaptive median filtering, calculates the extent of filtered image grads by using an adaptive method of eight-dot adjacent regions gradient, and automatically adjusts the high and low threshold values of the edge examination by using grads entropy and standard deviation according to the image edge grads informatioru Experiments show that, compared with the traditional Canny operator, this method does well in resisting image noises, keeping the detailed edges and the connectivity of edges so as to improve the automaticity of edge detection and achieves better practicability.

关 键 词：边缘检测 中值滤波 自适应阈值 CANNY算子 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]