基于图论的图像分割方法关键技术研究

KeyTechniques of Image Segmentation Based on Graph Theory

对基于图论的图像分割方法进行了相应的研究,针对传统分割方法不能通过算法自动实现对图像的分割,以及图像分割的质量与数据模型的建立有较大的联系,提出一种改进的基于图切割的自动分割算法,自动提取目标点和背景点,建立能量函数和网络图,实现图像的自动分割。实验表明,改进算法能够更有效地从背景中把目标物体分割出来。

In this paper, image segmentation methods based on graph theory was researched. In view that traditional method cannot automatically realize image segmentation and the quality of image segmentation was related to the established data model, an improved automatic segmentation algorithm based on graph cut was proposed, which could automatically extract target and background points and establish energy function and network graph, and realize automatic segmentation of image. It was shown by experiment that the improved algorithm could be more effective in segmenting target object from the background.