OpendTect User Documentation version 3.2
Prev		Next

GapDecon

Name

GapDecon -- Attribute that aims to attenuate repetitions of primary reflections (multiples).

Description

The type of multiple removal algorithm chosen for this application is the well known inversed filtering method also known as Gap deconvolution. This filter aims to attenuate a user-defined part of an auto-correlation function. The underlying idea is that multiples in the data are secondary reflections, i.e repetitions of the primary reflections that show up in the auto-correlation function at a time that corresponds with the extra travel time. The filter can be applied on-the-fly or in batch-mode to produce a filtered output cube. The user defines the GapDecon attribute from the list and specifies a number of input parameters;

Correlation window = the window for calculating the auto-correlation function.
Examine = pressing this button will open a new window where the user selects a subset of the data (e.g one inline, or one cross-line) on which the result of the auto-correlation will be calculated and displayed for filter design purposes. This result can be used to determine the lag and gap size.
Lag-size = the window length in the auto-correlation function that is unaffected by the filter. This window contains the wavelet-shape information.
Gap-size = the window length in the auto-correlation function that the filter aims to blank. This window contains repetitive (multiple) information.
Noise level = random noise that can be added to stabilize the filter derivation. The noise level is set to 1% by default.
Stepout = a moving average across neighbouring traces that is applied prior to the calculation of the auto-correlation function. This trace averaging results in a smoother end-result. The higher the stepout, the smoother the result.
The user can set the input and output to minimum or zero phase; if the input is zero phase, a phase rotation of 90 degrees is applied before applying the GapDecon filter. In case the output must be zero phase, the inverse 90 degrees rotation is applied after the filtering.
Check parameters = check whether the (multiple) energy has indeed been removed. A QC of the parameters : the GapDecon-filter with the selected parameters, lag 50 and gap 120 (see example above), is applied on a user defined line and the auto-correlation of the filtered data is displayed in a 2d-viewer to check whether the parameters are correct. If they are not correct, the parameters can still be changed.

Spiking Deconvolution. The Gap Decon attribute can also be used for Spiking Deconvolution or Whitening. The goal of Spiking Deconvolution is to flatten the output spectrum. This is achieved by shortening the embedded wavelet and attempting to make it as close as possible to a spike (zero-lag spike). One should keep in mind that the frequency bandwith of the data might limit the extent to which this whitening is possible. At higher frequencies, Spiking Deconvolution might cause an increase in noise.

Prev	Home	Next
Frequency Filter	Up	Horizon