基于万有引力搜索算法图像分割的实现

图像分割是计算机视觉中研究的热点和难点之一，原始图像只有在分割之后才能被分析与理解。为了更快更准确地分割图像，本文将万有引力搜索算法和最大类间方差法的双阈值法相结合，提出了一种新的图像分割方法。该方法首先把图像分割的阈值看成引力搜索算法中空间的粒子；其次利用最大类间方差法的双阈值法设计适应度函数；最后，通过粒子在空间中相互的万有引力作用，逐渐逼近最优阈值。实验表明，基于万有引力搜索算法(GSA)的图像分割在运行速度、运行时间等方面要优于传统的图像分割算法，该方法分割后的目标图像更加适合后续的分析和处理。

Image segmentation based on the universal gravitation search algorithm

Image segmentation is a hot and difficult problem in computer vision research, and the original image can be analyzed and understood only after the image segmentation. In order to segment images faster and more accurately, a new image segmentation method is proposed by combining dual threshold algorithm and Otsu gravitational search. Firstly, the image segmentation threshold is taken as the particle in space in Gravitational Search Algorithm(GSA). Secondly, the fitness function is designed by using dual threshold method of the Otsu. Finally, the optimal threshold is gradually approached through gravitational interaction between the particles in space. Experimental results show that GSA is superior to the traditional image segmentation methods at the segmentation speed and the running time; the image segmented by the proposed method is more suitable for the analysis and subsequent processing.