MAXIMUM LIKELIHOOD FILTERING OF REFLECTION SEISMOGRAPH DATA

Two types of adaptive multichannel filters applied to exploration reflection seismograph data are the maximum likelihood filter and the Wiener filter. Derived equations for the two types of filters demonstrate that the maximum likelihood filter equations are identical in form to the Wiener equations and are solved by the same matrix inversion techniques. Application of the maximum likelihood filter to common‐depth‐point (CDP) data demonstrates that the output reflection quality is sensitive to reflection alignment on input traces. Presently, sufficient coherency of reflections on input traces can be achieved only on CDP traces with a time window in which the reflection signal is dominant or on vertical array data for which reflection alignment is straight‐forward.