基于蛙跳算法与Otsu法的图像多阈值分割技术

Multilevel thresholding segmentation based on shuffled frog leaping algorithm and Otsu method

为了快速准确地确定多阈值图像分割中的最佳阈值,提出了一种基于蛙跳算法与Otsu法相结合的多阈值图像分割方法.该方法将多阈值求解看作一种多变量的组合求解优化问题,利用多阈值Otsu法设计分割目标函数,将新兴的仿生学优化求解算法——蛙跳算法引入到图像分割技术中,通过蛙跳算法中全局搜索和局部搜索相结合的搜索机制并行求解多个阈值.实验结果表明,该方法与基于人工鱼群算法的图像多阈值分割方法相比,明显提高了图像分割速度和分割质量.

In order to obtain a group of satisfying thresholds in image segmentation quickly and accurately, this paper proposed a method based on shuffled frog leaping （SFL） algorithm and Otsu method for multilevel thresholding image segmentation. The method regarded the group of thresholds as a group of potential solutions to a certain objective function, and employed the extended Otsu method to be the fitness function for SFL algorithm. And then, the powerful searching ability of SFL algorithm was used to locate the thresholds in parallel, which combines the global search in the whole swarm and local searches in subswarms. Experimental results showed that compared with the method based on artificial fish swarm （AFS） algorithm, the suggested method obviously im- proved the performance of image segmentation in speed and quality.