Automatic Stratification Strategy of Well Logging Curves Based on Wavelet Transform

Based on the feature analysis of well logging signal noise, we try to establish a well logging noise model in this paper. The problems focus on the implementation of Symlet wavelet and cubic spline wavelet on curves stratification of well logging signal. According to the specialty of well logging signals, we adopt discrete wavelet transform and wavelet packet to be used to decompose the logging signal, which can saves vertical resolution of logging curve, and maximum possible recovery of thin layer information. The simulations indicate the wavelet analysis method for automatic logging curve stratification proposed in this article better reflects the interface of lithology change and the result of automatic stratification is basically consistent with the actual situation, which shows obvious theoretical and practical significance to the research of well logging curves stratification.