论文部分内容阅读
通过制作理论模型，提取出时域、频域内最能反应薄层厚度的多种参数：反射波的总能量、视振幅、正半周面积、负半周面积、相关函数第１个零点的时移值、振幅谱面积、主频、中心频率、频带宽度、频谱的二阶矩及三瞬剖面等，它们与薄层厚度的关系都是非线性的。采用ＢＰ法神经网络，通过对模型数据的学习、记忆、识别，预测薄层厚度，取得了较满意的结果。
By making a theoretical model, we extract a variety of parameters that can reflect the thickness of the thin film in the time domain and frequency domain: the total energy of the reflected wave, the amplitude, the positive half-cycle area, the negative half-cycle area and the time-shift value of the first zero of the correlation function , Amplitude spectrum area, dominant frequency, center frequency, frequency bandwidth, second moment of spectrum and three-moment cross-section. All of them are non-linear with the thickness of thin layer. BP neural network is used to predict the thickness of thin layer by learning, remembering and recognizing the model data, and the satisfactory results are obtained.