Abstract: The morphological filter with omni-directional structuring elements is a typical nonlinear filter，which has a good performance for selecting direction edge information and removing the noise.It is widely used in image denoising and edge extraction.Based on general open maximum and general close minimum morphological filter，a new fast edge tracking algorithm is proposed.The key problems of algorithm are solved，by encoding edge pixel with edge direction information，calculating the minimum direction of the edge pixel tracking，and determining non-continuous point of the edge.In order to verify the effectiveness and feasibility of the algorithm，some contrast experiments with different algorithms are completed about the standard circle and the noised Lena image.Experimental results show that the new fast tracking algorithm has good features for anti-noise and single-pixel connectivity.

Key words: directional morphological filtering operator, edge tracking, nonlinear filter

摘要： 方向形态滤波器是一种典型的非线性滤波器，具有良好的方向边缘信息选择和去除噪声的性能，被广泛地应用于图像去噪和边缘提取。在广义形态开最大和广义形态闭最小的滤波器基础上，提出了基于多方向形态算子的快速边缘跟踪算法。主要解决了利用边缘方向信息进行边缘像素编码、边缘像素跟踪最小方向计算和非连续点判断方法等算法关键问题。为了验证该算法的有效性和可行性，对标准circle图像与加噪Lena图像进行了不同算法对比实验，实验结果表明新快速边缘跟踪算法具有较强的抗噪性和单像素连通性。

关键词: 方向形态滤波算子, 边缘跟踪, 非线性滤波器