基于自适应云粒子群算法的造影图像增强

Contrast Image Enhancement Based on Cloud Adaptive Particle Swarm Algorithm

针对造影图像增强现有算法的不足,采用自适应云粒子群算法。首先,由云模型产生云滴,确定云的特征值;接着正向云发生器产生粒子,基于粒子个体适应值把种群分3个云子群,且分别采用不同惯性权重的调整策略,自适应算子使权重随着粒子适应度值的减小而减小,从而实现了较优粒子取得较小的权重,在适应度最大时并非绝对的零值,从而提高了避免陷入局部最优的能力;最后,造影图像通过对数灰度变换增强,把求最佳的参数组合问题转化为自适应云粒子群算法寻优问题。实验仿真证明,在云隶属度大的地方图像像素的灰度域比较集中,增强效果边沿清晰。

According to the contrast image enhancement algorithms,adaptive cloud particle swarm algorithm is used. First of all,the cloud model produces cloud droplets, and determines cloud characteristic value. Then normal cloud generator generates particles based on the particle individual fitness, the population is divided into three cloud groups and used different inertia weight adjustment strategy, adaptive operator so that the weight with the parti- cle＇s fitness decreases,thereby realizing a particle made the smaller weight,in the largest fitness is not absolute zero, thereby improving the avoid local op- timum capacity. Finally, angiography images by logarithmic gray transform enhancement,to seek the best parameter combination is transformed into the problem of adaptive cloud particle swarm optimization problem. The experimental simulation results show that, in the cloud membership of local image pixel gray region is relatively concentrated,with clear edge enhancement effect.