基于边缘流和区域合并的岩屑颗粒图像分割

Image segmentation of cutting grains based on edge flow and region merging

边缘流分割算法可利用图像的多种特征进行准确的图像分割,但传统的边缘流分割算法运算复杂度高,容易造成过分割.针对这些问题,作者对边缘流算法进行改进,并提出一种基于边缘流和区域合并的图像分割方法.该方法首先对原始彩色图像进行改进的边缘流分割;再通过曲线演化和边缘连接得到封闭的边缘;最后根据区域颜色相似度对初分割的图像进行区域合并,得到最终的分割结果.实验表明,该方法提高了分割效率,解决了过分割问题,将该方法应用于岩屑颗粒图像分割取得了较好效果.

Edge flow segmentation algorithm can achieve precise image segmentation by using various features of an image. But the traditional edge flow segmentation algorithm requires a lot of computation- al time and results in over-segmented. To solve this problem, the way of constructing edge flow vector field has been improved, and a method for cutting grains segmentation is proposed based on edge flow and region merging. First, edge flow segmentation algorithm based on LAB model is applied to segment an original color image. Next, discontinuous edges are connected by curves evolving and edge linking. Finally, a region merging algorithm based on regional color similarity is performed to solve the over-seg- mented problem and then get the final segmented image. Experimental results show that the approach we proposed improves the efficiency and can solve the problem of over-segmentation, and it is efficient for segmentation of the cutting grains.