基于神经网络的光栅莫尔条纹误差模型参数优化

Parameter Optimization of Grating Moire Fringe Error Model Based on Neural Network

光栅莫尔条纹的误差模型是一种多峰的复杂正弦波动函数级数形式,单一的基于导数的非线性参数拟合方法可能无效,神经网络方法不依赖于函数导数,仿真实验表明模型参数的拟合符合实际。

The error model of grating moire fringe is a multi peak complex sine wave function in series form, simply based on the derivative of the nonlinear parameter fitting method may be invalid. Neural network method does not rely on the derivative of the function, the simulation experiments show that the fitting of model parameters are consistent with the actual.