Study Of Several De-noising Methods For Transient Electromagnetic Signal

The transient electromagnetic data measured from field is influenced by various noise and interference factors, so noise signals will cover up real available signals, the existence of noise leads to low exploration resolution when data is processed and interpreted, and can bring false information and affect exploration results. Hence, it is necessary to cover de-noising process to measured transient electromagnetic data before final signal process, on purpose of eliminating noise signal as well as extracting useful information of signal effectively. Then, a further step toward research ensues to make better use of transient electromagnetic data for image making and expound on geological condition. Accordingly, it is practically significant to apply appropriate de-noising means to increase the SNR of transient electromagnetic signal and extract available information in transient electromagnetic signal with noise.In practice, three index number approaching non-line method is often used to de-noise transient electromagnetic signal, but result isn't ideal. Wavelet analysis method is a kind of method in time domain and frequency domain with characteristics of window size constant, time and frequency window variable, at the same time, the method has the feature of multi-resolution analysis and is suitable for us to deal with unstable signal, especially dynamic,transient signal. Therefore, it is quite necessary to introduce wavelet analysis to transient electromagnetic exploring area to process transient electromagnetic signal aiming at improving SNR.Methods based on wavelet analysis are various. According to transient electromagnetic signal characters, the main research contents are as follows:1. Characteristics of the transient electromagnetic signal and conventional electromagnetic de-noising methods are analysed. This paper discusses three index number approaching non-line de-noising method, and processes measured transient electromagnetic signal by conventional de-noising methods;2. This paper studies detailed wavelet transform theory, including the continuous wavelet transform, discrete wavelet transform, the binary scale wavelet transform, and so on;3. Method based on signal de-noising of wavelet threshold value is studied, Then, this paper gives several forms of threshold, wavelet threshold de-noising steps, three kinds of wavelet threshold processing methods. By calculating, many parameters are established, such as optimum decomposed number of layers, choice of threshold value, choice of wavelet, and so on. It achieves to de-noise to associated noises of measured transient electromagnetic signals and SNR can increase from 5.5956dB to 6.8491dB. SNR of data increase with wavelet decomposed number of layers. The more similar shapes of signal and wavelet are, the better the effect of de-noising is.4. Method based on wavelet transform mold maximum value de-noising is studied, the Lipschitz index is analysed, the wavelet transform mold maximum value of signal and noise in the propagation characteristics of different scales are discussed. This paper studies the wavelet transform mold maximum value de-noising principle and de-noising algorithm. According to the turn signal projection reconstruction algorithm, the theoretical model and the measured transient electromagnetic noise signal are processed, and SNR increases from 5.5956dB to 6.8491dB. With the method, this paper wipes off associated noises of measured transient electromagnetic signals and gets a smooth data curve, but in dramatic changes of the region of the measured transient electromagnetic signals reconstruction will have a greater error;5. Method based on wavelet packet signal de-noising is researched. Then, this paper gives wavelet packet space decomposition, wavelet packet algorithms, wavelet packet de-noising steps. It achieves to wipe off associated noises of measured transient electromagnetic signals, and SNR can increase from 6.8092dB to 7.6011dB. Wavelet packet de-noising can not only deal with high frequency part but also low frequency part;6. The advantage together with disadvantage and application field of several de-noising methods are discussed, and validity of the methods is elaborately discussed, the fact that better effect of vision can be acquired by applying wavelet analysis in transient electromagnetic signals de-noising is well testified, wavelet analysis increases the SNR of transient electromagnetic signals, and provides theoretical basis for transient electromagnetic data imaging and geological state explanation. On the basis of above research work, around "Based on the transient electromagnetic data de-noising method of wavelet analysis" can outspread research. Aiming at the diversity and complexity of electromagnetic noise and the different applications characteristics of various methods, it can use a variety of integration methods for de-noising. Some problems of wavelet threshold value de-noising method are studies, such as choice of threshold value, choice of wavelet, and so on. The choice of these parameters directly relates to the quality of de-noising. The parameters are different in different application areas. It will be an important issue worth studying on how to choose the appropriate parameters in various fields to achieve the best de-noising results.