一种基于像素连接性的边缘图像分割方法

An edge image segmentation method based on pixels connectivity

针对边缘图像中有效对象的分割,提出一种基于像素连接性的边缘图像分割新方法。该方法首先根据对原始灰度图像进行Canny算法边缘检测,依据边缘图像中有效对象一般具有相对较好像素连接性的特点,对二值边缘图像实施标记处理,选择标记矩阵中具有极大和较大像素连接性的部分,即获取了边缘图像中的有效对象,完成边缘图像中对有效对象的分割处理。人造测试图像和实际灰度图像的仿真实验均表明,此边缘图像的分割方法是有效的。

A new method, which based on edge image binary pixels connectivity, is proposed to implement objects segmentation of edge image. While achieve this method, there is a Canny edge detection is applied at first, according to the facts, that the significant objects in edge image always have an effective connectivity, then a label processing is carried out for the mort greater or the maximum element numbers of the connected components,these pixels are component to these significant objects, this is the proposed segmentation method. Extensive simulations are carried out to validate the proposed method. Experimental results show this method is effective.