[vtkusers] Data format to use for 2D image data in a 3D sheet/curtain

Brendan Smithyman smithyman at gmail.com
Mon May 13 16:05:16 EDT 2013


Hi All,

I'm trying to visualize "2D" seismic data using Paraview.  This means
that the dataset has 2D connectivity (as in a cross-section image), but
appears as a sheet or curtain in 3D space with the data traces running
vertically.  The vertical spacing is constant, however, the top of each
vertical trace can be located arbitrarily in 3D (but for argument's
sake, let's say that the z-axis location of the top of the trace is
flat, which is often the case).  I've written some Python code to
convert from the SEG-Y format to a vtkStructuredGrid XML file (linked
below), but I'm not sure this is the right format to use.  In
particular, it would be nice to set the mesh connectivity up such that
adjacent data points are connected in a surface, i.e., both along the
curved line and vertically.  At present, the implicit connectivity seems
to cause the bottom of one trace to connect to the top of the next,
which causes major issues if I want to work with a subset of the data.
In my case these files are maybe 3000 or 4000 traces wide by about 2000
samples high, so 6-8 megapixels without resampling (though 2D/3D seismic
data can be quite a bit bigger or smaller in terms of data size).

Question:
Do you have any ideas or recommendations as to the data format I should
be using?  I've put together some images so that you can see the issue.
The data in the image are resampled to every 10th point in both data
dimensions, so nominally 360x200 in the right-most segment.

Example images:
https://www.dropbox.com/sh/y5v20zl3vczz4a7/G7-cCqR9pR

Current code:
https://skylab.bitsmithy.net/svn/public/pygeo/branches/devel/util/stack2vtk.py


Many thanks,
Brendan Smithyman
PhD Candidate @ UBC Vancouver




More information about the vtkusers mailing list