一种基于Canny的边缘检测改进算法

摘要

针对传统Canny算子在边缘检测时需要人工设定高斯滤波参数以及高、低阈值等问题,提出一种基于Canny的改进边缘检测算法.首先,采用改进的自适应中值滤波对噪音图像进行平滑处理,在去噪的同时锐化边缘;其次,借鉴一阶Sobel算子模板,在其基础上拓展为水平、垂直、45°、135°四个方向的梯度模板计算梯度幅值;最后,结合图像灰度直方图统计自适应生成阈值.实验结果表明,改进算法较原始算法效果更优.%For the problem that the Gaussian filters parameters and the high and low thresholds in the traditional Canny algorithm should be manually determined,the modified edge detection algorithm based on Canny is proposed.Firstly,the modified adaptive median filter is used to smooth the noise image,and the edge is sharpened while denoising.Then the modified algorithm uses 4 gradient templates to calculate gradient amplitude in horizontal direction,vertical direction,45° direction,and 135° direction.Finally,the threshold of the image gray scale histogram is used to generate the threshold.Experimental results indicate that the modified algorithm is more effective than the original one.