Data-Adaptive, Robust, Non-Linear Algorithms For Noise Removal On Stationary Or Non—Stationary Time Or Space Series Data

Basically all geophysical data consist of readings made in a time and/or<br>space series manner. The general problem is to separate the noise<br>components from the signal components. To succeed, some knowledge of what<br>is noise and what is signal are necessary. Unfortunately most methods<br>break down when the data are not stationary and contain spurious peaks or<br>outliers. Sometimes no information of the noise can be obtained. In<br>this paper the results of some new algorithms will be shown that was<br>developed to tackle the problem in these circumstances. Four methods were<br>developed. The basic concepts of each method is outlined and examples on<br>synthetic and real data are given.