一种新的基于Graph cuts方法的SAR图像分割模型

Graph cuts方法是一种快速优化技术，能有效解决计算机视觉的低层次问题，如表面重构、分割、去噪等问题。其优点是迭代计算的高效性并能得到全局最优解。利用Graph cuts方法，将基于活动轮廓模型和水平集方法的SAR图像分割模型转化成一个新的模型，可用最大流/最小割方法来求解。实测SAR图像的分割实验表明：提出的新模型所需运行时间大约是基于水平集的活动轮廓模型所需时间的一半，分割精度也得到了较大的提高。

A New SAR Image Segmentation Model Based on Graph Cuts Method

Graph cuts is a fast and optimization technique, which can effectively solve the problem of low level computer vision, such as surface reconstruction, segmentation, and denoising problem and so on. Its advantages are efficient iterative calculation and can get the global optimal solution. Using Graph cuts method, SAR image segmentation model based on active contour model and level set method was transformed into a new model, which could be solved by max flow /min cut method. By segmenting envisat SAR images, the running time required for the proposed new model was about 1/2 of the running time required for active contour model based on level set method, and segmentation accuracy were greatly improved.