Computation of a fault property from the seismic data cube

@inproceedings{JacqueminRM2008,
 abstract = { Until now there is no good enough method to automatically detect the location of faults in a seismic data cube. For example one can use the “semblance” property, but faults can then hardly be separated from the noise. This implies that we must use the only reliable information given by the reflectors, and only this one. Therefore we propose computing the correlation value between a small window in a 2D seismic cross section and its neighborhood. From this neighborhood we keep the best correlated window. If the correlation value is high enough, it means that we are on the same layer. We can now consider the difference of z level between the center of this small window and the one of the original window. We are on an unbroken layer if the difference is small. On the contrary, in case of high z step, this means that we have found the other side of the same broken layer and we can say that there is a fault between the two windows. The advantage of this method is that we use only the reliable signal given by the reflectors and that we eliminate deceptive information given by ghost traces because there is no slip in that case. Also, this method produces the vertical fault drag, which is a valuable input information for structural modeling },
 author = { Jacquemin, Pierre },
 booktitle = { 28th gOcad Meeting },
 month = { "june" },
 publisher = { ASGA },
 title = { Computation of a fault property from the seismic data cube },
 year = { 2008 }
}