To remove the noises of EEG effectively, this paper makes the EEG De-noising research about Wavelet and Hilbert Transform. In HHT De-noising process, first, according to EEG own frequency characteristics, the EEG signals are made eight scales decomposition by using EMD algorithm, and obtain eight IMF component signals. Second, reconstruct the IMF component signals after filtering. Finally, get the EEG after De-noising. The experimental results show that HHT method can preferably eliminate the noises which mixed in the EEG. The De-noising effects of HHT and Wavelet Transform methods are compared by using the evaluation indexes. It finds that HHT method is superior to the traditional Wavelet Transform in the EEG De-noising, and its efficiency is higher.

The correlation method had once been considered as one of the best methods for the measurement of multiphase flow. However, if the behavior of flow does not fit the ergodic random process, the measured cross correlation plot will have a gross distortion when the different components of flow do not pervade within one another to the full extent. We measured a variety of parameters of three phase oil/water/gas flow in an oil pipeline. The change of flow pattern is so complex that the measured signals are always contaminated by stochastic noises. The weak signals are very easily covered by the noise so that it will result in great deviation. Wavelet transformation is an analytical method of both time and frequency domain. The method can achieve signal decomposition and location in time and frequency domain through adjustment and translation of scale. An LMS algorithm in wavelet transform is studied for denoising the signals based on the use of a novel smart capacitive sensor to measure three phase oil/water/gas flow in oil pipeline. The results of simulation and data processing by MATLAB reveal that wavelet analysis has better denoising effects for online measurement of crude oils with high measurement precision and a wide application range.