针对图像增强的特点,采用云编码算法对策。不定长的自然数编码机制和基于顺序云滴个体的适应度函数搜索云滴最优解;采用赌轮选择法产生新一代云滴,用自适应的交叉概率求解相对某个解的最佳交叉概率,在保持云群体多样性的同时,保证算法的收敛;通过整体目标函数淘汰适应度小的个体。实验仿真显示云编码算法对图像增强效果最好,处理时间少,信噪比大。

In order to improve the quality of image enhancement, cloud code algorithm was proposed. Firstly, uncertainty of natural number code was used in the optimal solution of cloud. The individual fitness function of cloud was based on the sequence. Secondly, the new generation of cloud droplet was produced by a roulette wheel selection method. The adaptive probability of crossover was provided by the best optimal and then kept population diversity and convergence of the algorithm. Finally, the fitness of individual was eliminated by the function of overall goals. The simulation shows this algorithm effect of image enhancement is the best, time-saving, and better signal-to-noise ratio (SNR).