一种边界-区域相结合的图像分割算法

Image Segmentation Algorithm Based on Region Segmentation and Edge Detection

提出了一种模仿人类视觉机制的区域-细节的图像分割算法。首先提取图像边缘,然后将边缘分段切割,得到端点集合,然后从端点集合生成Delaunay三角形网络,以Delaunay三角形为顶点,相邻三角形的属性差异作为边的权重,构造图;以基于图的分割算法生成最小生成树,划分区域。最后用Snake模型精确确定区域边界,生成准确的区域边缘。实验证明,这种区域分割和边缘检测相结合的方法能准确地分割非纹理图像,较好地克服了块现象和非连续边界,相比单一区域分割或者边缘检测方法有更好的分割结果,并且计算速度比较快。

An approach for image segmentation is introduced based on region segmentation and edge detection. Edges are first detected in the original image and then cut equidistantly into some segmentations, and the end points of segmentations are gathered into a points set. A constrained Delaunay triangulation is yielded from the points set and then a graph is attained too. Graph cut, Minimum Spanning Tree （MST）, is performed. Finally, the obtained regions are projected into the original image and region boundaries are refined to achieve pixel accuracy by Snake model. Experimental results show that the approach can achieve accurate edge localization, better spatial coherence and improved efficiency.