一种基于方形函数的邻域灰度差值边缘提取方法

Method of Neighborhood Grayscale Difference Edge Extraction Based on Square Function

根据数字图像处理中现有边缘检测算子,对连续边缘的检测及抵抗噪声的能力问题,提出一种基于方形函数的测算法。首先将目标图像划分成方形的采样网格;然后确定检测的起始点位置,利用图像边缘处内外灰度差值最大原理判定该方形区域边缘所在的切线位置;最后通过多个区域的并集取得图形边缘。实验证明通过该算法可以提取出较为完整的边缘,抗噪能力也优于以往的边缘检测方法。

According to the ability of continuous edge detection and noise resistance of existing operators, a new edge detection operator based on square function is proposed. First, the target image is divided into square sampling grids. Then the position of start point is determined,and the tangent position is found out maximum by the gray scale differ-ence at the edge of the inner and outer areas. Finally,the edge of the target is obtained by integrating the same areas. Experiments show that a rather complete edge can be extracted in this way and the noise resistance ability seems stron-ger.