摘要
利用经典的Otsu算法和基本遗传算法相结合进行图像分割存在有算法效率低、容易提前形成伪解的问题,对于上述问题,提出一种基于改进小生境遗传算法的图像分割算法(IVNGAMS)。算法全局优化了二维Otsu图像分割函数,可以按照个体适应度大小自动控制遗传参数。并通过引入模拟退火算法,进一步提升算法的局部搜索能力。实验结果表明,改进的图像分割方法能更好提升算法的全局搜索能力,能够更加稳定快速的收敛到最佳的分割阈值,并且得到了更好的图像分割效果。
The classical Otsu algorithm and the basic genetic algorithm combined with image segmentation have the problem of low efficiency and easy to form pseudo solution in advance. For this problem, an image segmentation algorithm based on improved niche genetic algorithm （IVNGAMS） is proposed. The algorithm optimizes the two-dimensional Otsu image segmentation function globally, and automatically adjusts the genetic parameters according to the individual fitness. And the simulated annealing algorithm is introduced to further improve the local search ability. The experimental results show that the improved image segmentation method can improve the global search ability, and can converge to the optimal segmentation threshold more stably and quickly, and obtain better image segmentation effect.
出处
《计算机应用与软件》
2017年第4期202-206,248,共6页
Computer Applications and Software
基金
国家自然科学基金项目(61363002)
广西教育厅科研基金项目(LX2014002)
关键词
遗传算法
小生境
图像分割
阈值
Genetic algorithm
Niche
Image segmentation
Threshold
