2019 · 12 citations · 6 references
Natural SciencesEarth SciencesGeophysicsApplied GeophysicsExploration GeophysicsTime Series AnalysisDigital Signal ProcessingGeophysical Data ProcessingGeophysical Signal ProcessingPulse ProcessingSignal ProcessingData SparsityDictionary TheoryGeophysical ModellingModel IdentificationMorphological StructureSound SignalsSparse Signal ProcessingSparse Signal Representation
The study of geophysical signals involves the analysis of their internal morphological structure. The signals of geoacoustic and electromagnetic emission are regarded in the paper. These are pulse signals and for their analysis it is required to use time-frequency methods. The paper presents the improvement of methods for sparse model identification of pulsed geophysical signals based on the use of parametric dictionaries. The dictionaries used in the sparse approximation should contain atoms which are similar to the signal morphological structures and at the same time be representative enough to reflect the different variations of the morphological structures. A comparison of various parametric dictionaries (Gauss, Ricker, Berlage, and Lognormal) is described in the paper. The choice of these functions is motivated by experience of their use in the geophysical pulse modeling. A modified adaptive matching pursuit (AMP) algorithm is used for the sparse model identification. The AMP allows us to expand the range of variable parameters of the dictionary functions. It is shown that the dictionary consisting of low-chirped Berlage pulses provides minimum sparsity for fixed error value for geoacoustic and electromagnetic signals.
6
Multichannel matching pursuit for seismic trace decomposition
Yanghua Wang · Geophysics · 2010 · 113 citations