WaveFormClass

Surface with learning points

The point dataset corresponds to the exact position, in the input seismic cube, where the program will take the various class typical waveforms. This option allows for conducting a semi-controlled classification. Usually, the well location points are taken so that the reservoir parameters are known.

Plug-in Overview

The WaveFormClass plug-in divides signal characteristics within a chosen interval into a number of classes of similar signal shape. A map of these classes enables performing facies analysis, exploring waveform changes in a lateral direction and predicting solutions for problems, based on the seismic waveform. Maps of probabilities for every class type can help estimate the quality and credibility of the classification.

Specifications

• Plug-in calculates new attribute for specified surface (top surface) with classes (seismic facies).
• Plug-in calculates new attribute for specified surface with probabilities for every class.
• Plug-in calculates objects that describe the waveform under the specified surface according to the parameters.
• Plug-in calculates attributes for specified surface that represent seismic amplitude in several points vertically.
• Plug-in calculates subcubes that represent typical 3D waveform corresponding to classes and situated in point where function probability for the class will be maximized.
• Plug-in allows using learning points, which additionally represents the locations for master wavelets.
• Plug-in allows using predefined waveform (list of wavelets), which can be used for classification.

Features

The first stage is the learning phase. During this stage, the typical waveform (2D or 3D waveform — sub-cube) for every class is defined. The second stage is the actual object classification phase. During this stage, the program calculates the similarity or probability value, which is a measure for the resemblance with the master trace for every seismic trace. This is done for each master trace of the various categories. The highest probability range is used to define the class category for the trace under examination.