摘要
提出了基于遗传算法的干涉条纹数据处理方法。该方法是在最小二乘余弦曲线拟合的基础上 ,引入了具有全局优化能力的遗传算法以解决目标函数多极值问题 ,从而使得最小二乘余弦曲线拟合对干涉条纹处理具有很好的鲁棒性和自适应能力。仿真结果表明 ,在 2 0 %均匀分布噪声条件下 ,拟合模型主要参数 -干涉条纹空间频率及初相位 -拟合误差小于 0 .5 %。
A novel data processing method of interference stripes based on genetic algorithms is described. In the course of curve fitting for interference stripes with cosinus, there usually exist several extrema of the objective function in the parameter space, which baffles the conventional optimization technique to converge to the right parameters. Therefore genetic algorithm is introduced to assume the optimization task. And the simulating results show that the fitting errors of the main parameters are less than 0.5 percent even in the existence of uniformly distributed noise, the amplitude of which is up to 10 percent.
出处
《光学技术》
EI
CAS
CSCD
2001年第2期136-138,共3页
Optical Technique
