阶段变异萤火虫群算法及在DR图像增强中应用

Section Mutation Glowworm Swarm Algorithm and Application in DR Image Enhancement

针对基本萤火虫群算法收敛速度慢、求解精度低的缺点,提出一种阶段变异方法来改进基本萤火虫群算法,即阶段变异萤火虫群算法。医学DR图像存在对比度低、边缘模糊等问题,对其增强以分段灰度变换为模型,6个分段控制参数为寻优变量,提出一种递归Otsu分割算法对寻优变量进行预设修正,采用阶段变异萤火虫群算法求解寻优变量值。实验结果表明,该增强方法具有自适应性,可以提高图像的分辨率和对比度,使图像细节更加丰富、清晰。

According to the problem that the low precision and convergence speed of basic glowworm swarm algorithm, and section mutation is put up, it can improve basic glowworm swarm algorithm, so section mutation glowworm swarm algorithm is put forward. There are problems with the low contrast and hazy border to medical DR image. To solve this problem, section gray transform is selected as enhancement model, six section control parameters are optimization variables, optimization variables are updated by recursive Otsu segmentation algorithm in advance, optimization variables are found by section mutation glowworm swarm algorithm. Simulation experiment on this method shows that it is automatic adaption,resolution ratio and contrast of image are enhanced,details of image is abundant and clear.