In seismic signal analysis, points of sharp variation classified as "edges" contain a considerable amount of a signal's information, thus making edge detection and the study of a signal's local properties an appropriate mechanism for obtaining information from seismic data. Several important physical processes can in principle affect the local regularity of a reflected event in a seismic trace: processes of absorption and wave attenuation. The local regularity of a given signal is characterised by the continuous wavelet transform and subsequently measured by its corresponding Lipschitz exponent(s). For a single seismic event resembling a delta type function, a linear model can be used in order to estimate the associated Lipschitz regularity, however for practical settings a non-linear objective function would have to be minimised in order to estimate the associated regularity. A robust estimation of a functions local properties and differentiability from seismic data, alongside prior geological information, could potentially lead to processing and inversion algorithms able to discern and characterise such targets.
View full article as PDF (1.09 Mb)