Application of wavelet analysis on thin-film wideband monitoring system

In the thin-film thickness wideband monitoring system, an accurate measurement of the spectral intensity signal is critical to improving precision of coating. Because of electron gun, ion source and baking, vacuum chamber is a complex environment with background light. Together with the inherent noise of linear CCD and quantization noise of A/D conversion, which are main factors affecting accurate measurement of spectrum intensity. This paper uses time and frequency multi-resolution properties of wavelet transform, adaptive threshold adjustment method is designed. According to the different characteristics of signal and the random noise processed by wavelet transform in different scales, the fine adjustment factor is added when the threshold is determined, on the hand, which makes the adaptive threshold of wavelet coefficient with positive Lipschitz index decrease, this is beneficial to preserve real signals of wavelet coefficient; on the other hand, which makes the one with negative Lipschitz index increase, this is favorable to filter out noise signal. By this method, both rejecting true probability and false declaration probability are reduced, the random noise is suppressed effectively, a very good filtering result is achieved, and finally the analysis accuracy of spectrum signal and the precision of systematic decision pause are improved.