摘要
结合遗传算法和模拟退火算法的特点提出整体退火遗传算法 (GASA)对相息图位相分布进行设计 ,从而达到降低噪声、提高再现像质的目的。该混合算法充分利用了GA的并行性及保留一定历史信息的特点 ,用SA的温度参数控制收敛性。利用空间光调制器作为记录装置 ,实验结果表明 :对于 6 4× 6 4像素的相息图 ,位相量化级达 16时 ,优化结果较理想 ,实验结果与计算机模拟再现像吻合得非常好。与传统模拟算法计算结果比较 ,在相同条件下 ,衍射效率和信噪比都有所提高 。
A novel design for phase distribution of kinoform by general annealing genetic algorithm combining genetic algorithm with simulated annealing algorithm is proposed so as to reduce the noise and improve the quality of reconstructed image of kinoform. GA,which is a paralle algorithm,reserve historical information for SA,whileSA's parameter,temperature,is used to control GA's convergence. The optimized kinoform is synthesized with a liquid-crystal spatial light modulator(LCSLM)as the recording device, experimental results show that the quatization level of L=16 for the phase consequent is sufficient for a 64×64 kinoform in the present optimization. The reconstructed images are agree well with the computer simulations. The signal noise ratio and the diffraction efficiency are higher than those of conventional method under the same condition.
出处
《应用激光》
CSCD
北大核心
2003年第4期225-228,共4页
Applied Laser
