差分变异本质粒子群的模糊熵图像分割

Image segmentation based on differential mutation bare bones particle swarm optimization and fuzzy entropy

基本本质粒子群算法存在易陷入局部最优以及过早收敛的缺点。在基本本质粒子群算法的基础上,借鉴差分进化中利用差分量对种群进行变异操作的思想,提出了差分变异本质粒子群优化算法。结合图像模糊熵,得到了基于差分变异粒子群优化的模糊熵图像分割算法。算法利用差分变异本质粒子群来搜索使图像模糊熵最大的参数值,得到分割阈值对图像进行分割。通过与其它两种本质粒子群算法的分割结果比较表明该算法取得了令人满意的分割结果,算法运算时间很小,能够满足对煤尘浓度实时精确测量的要求。

Basic bare bones particle swarm optimization （BBPSO） is easy to get stuck into local optima. Based on basic BBPSO, using the idea of mutation in differential evolution, a new algorithm named differential mutation bare bones particle swarm optimization （DMBBPSO） is proposed and combined with image fuzzy entropy to obtain a new segmentation algorithm based on DMBBPSO and fuzzy entropy for image segmentation. The proposed algorithm uses DMEBBPSO to explore fuzzy parameters of maximum fuzzy entropy and gets the image segmentation threshold. According to the experiment results, compared with other two algorithms, the proposed algorithm shows better segmentation performance and very low time cost. It can be used to real time and precision measure coal dust image.