Empirical Mode Decomposition versus Wavelet filtering for Time Domain Electromagnetic Signals

Abstract

Time Domain Electromagnetic (TDEM) methods are used for exploration of deep seated mineral deposits by measuring sub-surface conductivity of the earth. The response of earth is measured by an induction coil. The signals acquired are contaminated by variety of noises. Signal filtering plays a major role in order to obtain deeper information regarding subsurface conductivities. Various methods based on conventional DSP filters, wavelets, etc. have been attempted. This paper explains in short a methodology of Empirical mode decomposition by intrinsic mode functions. The potential of this technique for filtering TDEM signal is demonstrated, further a comparison of EMD technique with wavelet based filter is illustrated.