Çatalhöyük Preliminary Interp

Çatalhöyük GPR Interp
Çatalhöyük GPR Interp

First crude stab at interpreting one timeslice of the 500 MHz GPR data collected last June at Çatalhöyük.

I had to prepare a couple of images this last weekend for an abstract for the London Geological Society Meeting in December and in the process came up with the first stab at digitizing some interpretations to one of the many timeslices from the 500 MHz GPR data collected last June. (More information on the survey can be found in a previous post) Initially I wasn’t very keen on attempting interpreting the slice as we still have not corrected the data for topography or tilt yet, but in the end I think it was a good exercise for me in coming to terms with actually making interpretations on the data- versus simply digitizing every reflection on the timeslice. I think there’s certainly some promising anomalies/features appearing in this particular slice (thus far estimated at 85cm below ground surface, but still a contentious estimate).

In addition to topographic correction and velocity analysis, there still needs to be quite a bit of mosaic correction done as well. As the survey was completed over the course of an afternoon and the following morning the relative amplitude ranges as well as the time zero points vary significantly between the two days. For the initial timeslices I ended up correcting the relative amplitudes through various attempts at adjusting the gain curves on each trace, automating and clipping each radargram to their respective time zero point, and then crudely adjusting the resulting amplitude ranges in the pixel based timeslices themselves. I think with greater refinement to the relative gain curves I can probably avoid the pixel based mosaic corrections, but have yet to have the time (and access to the software) to have a real play with the data. Soon enough though- as the goal is now to make some real progress in preparation for the poster in December. Here’s hoping!


4 thoughts on “Çatalhöyük Preliminary Interp

  1. Hi Jess,

    OK, you know I know nothing about geofizz … but …

    I’m looking at your 2 images and see linear features in the left view that don’t appear in the right (did you get the image on the right by thresholding?). I’m wondering about using a gradient filter on the data – I expect you’d still keep these lines that way. Or a density slice to pick up those mid greys/low signal contrast features? That could then be added to an image with a lower threshold for high contrast features. Now I’m thinking a Posterize function where you could control not just the number of levels but their ranges too. Now that I’ve frothed on way too much, you can tell me that all of this is just basic stuff to the Geofizzysist, that you did it this way for good reason, and go blow it out my ass 🙂

    Cheers, Geoff

  2. Thanks for your comments Geoff! “Preliminary” is the key word here! I guess the first thing- which I probably should have made clear is that this interp is not the result of any automated extracting processes, but rather the result of me manually interpreting and digitizing interpretive polygons in the GIS. It is by no means all encompassing, and I’m increasingly aware (the more I look at it) that there are missing features, and particularly that the low amplitude, less contrasting features are not well represented here. My next goal (as I mentioned above) is to have a go at topographic correction, and some more stabs at adjusting the relative amplitudes between lines. Not sure if you’ve looked, but I posted a video animation of the full data set which scrolls between the timeslices. It’s abundantly clear in this data set that there is striping and edge-matching issues and that additional mosaic corrections are certainly required. Also, it should be mentioned that I did do quite a bit of image processing to produce the slices (though always concious that in GPR, it’s very easy to filter away features of interest..), including high and low pass filters. Some of those faint linears that you are seeing are somewhat stronger in other slices at different depths, too. At the end of the day, this sort of method for interpreting GPR data is crude, and really does not suit the 3D nature of GPR data. I do have some intentions to try and play with automating the process though, through supervised classifications and voxel-based models, so will keep you posted!

  3. So many questions to ask but I expect that in doing so, I’d simply be demonstrating my ignorance of the subject and taking too much of your time on what I expect is geofizz basics. So maybe I’ll collar you at CAA for a long chat over a drink (you’ve been warned :). But I’m definitely liking the use of volumetric models (voxels).

    Catch you in April (if not before),

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s