USGS Digital Data Formats

April 19, 2004 6:55 PM

Is anyone familiar with old USGS Digital Data Formats?

I have some old USGS image libraries on CD-ROM that I'd like to salvage from proprietary data format land. The CD's require a DOS program called GEOPH.EXE to browse, (which comes on the CD). The images are just PCX's, which my new fangled computer can read without a problem.

The image descriptions, however, are binary files with an ABS extension. I don't think that ABS refers to a specific file type, but rather a general science abstract. Here's an example of one of the files.

The CD-ROMs were published in 1995.

Is anyone familiar with the data format and how it might be read/decoded? The GEOPH.EXE program will let you export photograph descriptions to a text file, but you have to open each picture (one at a time) to do so, and I'm not that interested.

Any help would be appreciated.

posted by alan to technology (8 comments total)

Interesting, the .ABS files are PCX files too, though the image data in them isn't in any way meaningful.

AAK0125.ABS: PCX ver. 3.0 image data bounding box [0, 0] - [1374, 1065], 8-bit colour, 200 x 200 dpi, RLE compressed.
posted by fvw at 7:40 PM on April 19, 2004

The GEOPH.EXE program will let you export photograph descriptions to a text file, but you have to open each picture (one at a time) to do so, and I'm not that interested.

A batch file (.bat) is what you need, then.
posted by trondant at 8:09 PM on April 19, 2004

Not that simple trondant. GEOPH.EXE is a DOS based image browser, which appears to be specifically designed for these CDs. To export the data you need to

1. Run the DOS program
2. Broswe a set of images
3. Press F9
4. Type in a file name
5. Select the image

In addition to viewing the image, GEOPH.EXE will append the descriptive information the to text file you specified. If you can come up with a way to batch that, you shouldn't be wasting so much time in ask metafilter :)

fvw, what did you use to extract that information?
posted by alan at 8:47 PM on April 19, 2004

I used file, a unix program that has come up here on AskMeFi before. It's in most BSD and linux distributions, you can also find it here if you want to run it under cygwin or something. A very useful utility indeed. I'm afraid I'm out of suggestions for decoding the file though, short of asking the USGS or reverse-engineering GEOPH.EXE. Are you sure the description data is in the ABS file?
posted by fvw at 9:16 AM on April 20, 2004

I ran strings on the sample file and I didn't see any ascii text. It's possible that the meta data is encoded or compressed somehow but I don't think that's likely to be the case. I think fvw is pointing in the right direction. Maybe the meta data is in a separate file somewhere. There are two things I would try here. First, run strings on geoph.exe to try to pick up a file name. Second, get sample meta data for a file and then grep through the file tree looking for files that contain the meta data. [All suggestions assume a unix environment]
posted by rdr at 12:15 PM on April 20, 2004

rdr: The file linked _is_ a separate file. For starters, it's much too small to contain any reasonably sized image. Secondly, as noted, it's a PCX so strings won't help you much.
posted by fvw at 1:48 PM on April 20, 2004

Short version. There are 10 files on the CD named

DDS21DEX
DDS21DEX.1ST
DDS21DEX.2ND
DDS21DEX.3RD
DDS21DEX.4RD
DDS21DEX.5TH
DDS21DEX.FAM
DDS21DEX.LOC
DDS21DEX.YR
DDS21TEX

that contain all the meta-data for the photographs in good old ASCII text.

Long Version. There are 4 CDs in the set, with the PCX's spread out across all the CDs. For each PCX there's a corresponding ABS file. However, each CD appeared to contain all the ABS files for images on the other CDs. One of GEOPH.EXE's features was a keyword search that spanned all 4 CDs (which would require all the meta-data be on each CD or installed locally). When I did a search for .abs as a file extension, I kept finding references to "scientific abstracts".

All of this led me to believe the meta-data was stored in those ABS files.

Which was wrong.

After fvw revealed the abs files to be nothing more than small PCX's, I started to think that the meta-data wasn't in there, and that the ABS's were just another bizarre piece of a DOS multimedia program.

That led me to do what I should have done in the first place, and grep through the entire CD directory, which turned up the above files.

Thanks to everyone to busting me out of my tunnel vision.
posted by alan at 3:38 PM on April 20, 2004

Which still leaves the question "What was in the ABS file?". Is the GIMP's (an image editor) PCX decoder off? It definately didn't look like a normal image…
posted by fvw at 5:38 PM on April 20, 2004