结合粒子群算法优化归一割的图像阈值分割方法

摘要

In order to get the optimal threshold in image segmentation quickly .based on the graph theory, gray-scale similar matrix takes the place of pixel-level weight matrix, normalized cut criterion is regarded as the optimization function. Using particle swarm optimization algorithm to find the best threshold in gray-scale space. Experiments show that the method is not only less computational costs, but also get a satisfactory segmentation result. The thresholds is more stable and consume less time greatly and better satisfies the request of real-time processing in image segmentation by using this new method.%为了快速得到图像分割的最佳阈值,依据图论知识,利用灰度级相似矩阵代替像素级权值矩阵,将归一化切割准则作为优化函数.利用粒子群优化算法代替穷举法优化归一化划分准则,提出粒子群算法优化归一割的图像阈值分割方法.实验表明在分割性能上有较大的提高,在分割速度上也有较大的改进,能够满足实时性要求.