Logging data preprocessing with combination of adaptive median filtering and wavelet transform

Due to the influence of conditions of logging, the random disturbance is superposed to the logging data, which has s nothing to do with the formation properties. The interference may be pulse noise can also be Gaussian white noise. For erasing the noise, the adaptive median filter is used for the impulse noise and the wavelet filter follows for the Gaussian white noise. Experimental results show that the combination of adaptive median filtering and wavelet transform can effectively filter random disturbance with the logging signal. It is better than simply using wavelet transform or median filtering method. In the application of classification of Strata the method achieved good result.