Thalweg tracker - OpendTect Pro feature A few weeks ago, we posted a new workflow for seismic facies classification. This week we show another example of OpendTect’s Thalweg tra...
Thalweg tracker - OpendTect Pro feature
A few weeks ago, we posted a new workflow for seismic facies classification. This week we show another example of OpendTect’s Thalweg tracker / Machine Learning workflow. In this case we visualized patterns in the Tubbergen Formation in eastern Netherlands. The Late Carboniferous Tubbergen consists of alternating sandstones and claystones / shales and rare coal beds with variable extent and thickness. Finding clean, thick, and extensive sands is commercially interesting as the Tubbergen has potential as geothermal energy source. We used the Thalweg tracker to create 10 pointsets with 10.000 locations each to serve as labeled class positions for a Convolutional Neural Network. All classes were tracked from starting positions selected at seismic-to-well calibrated gamma-ray logs; 6 from the well in the forefront and 4 from the well in the background. The slider shows the input seismic flattened at Base Zechstein, the input pointsets, and the Machine Learning classification result in the mid Tubbergen.
#opendtect
