论文部分内容阅读
针对莫尔条纹信号误差不能满足高精度的测量缺点,提出一种基于RBF神经网络方法,采用样本点的正切值作为网络的输入,将训练样本点分段学习的方法对网络进行训练,仿真实验表明:该方法计算出的理论误差与实际值吻合较好。
Aiming at the shortcomings that the error of the moiré fringe signal can not meet the requirement of high accuracy, a method based on RBF neural network is proposed. The tangent value of the sample point is used as the input of the network, and the training sample point is segmented to learn the network. It shows that the theoretical error calculated by this method is in good agreement with the actual value.