|The Department Graduate Students Spacecraft Missions|
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 14 | 15 | 16
Paul Geissler, on the founding of PIRL
Image processing was really just getting started. It was very awkward, very hard to do, so it was a good time to get into it. They had hired a new professor named Bob Singer, and I wrote to him even before he showed up and said that I wanted to work with him. I basically pounced on him before he even unpacked, and he agreed to take me on as a graduate student.
His idea was to start a Planetary Image Research Laboratory [PIRL], which still exists. It’s the one that Alfred [McEwen] is head of now. Bob started it. Brad Castalia was one of the first principal programmers there.
Basically it started off with eight single Sun workstations and nine-track tape drives because that’s where we got images from. We would get images and download them on the tape and basically write software to be able to process them.
Our recording medium was a camera and a tripod that took a photograph of the screen to make a slide; it was very basic. We had a gorgeous monitor—it cost twenty thousand dollars in the late 1980s. It had 1024 by 1024 pixel resolution: Unheard-of.
We picked up equipment as we went. We eventually got a digital film recorder. We got a lot more workstations since they got cheaper as more graduates started working on images. But for a while it was really just a handful of us that were trying to process images. Normally what would happen is people would get hardcopy, photographic hardcopy, and they would slice it up and paste it together.
The things that we did were actually pretty impressive, because we didn’t really have any software that we could run. We made a decision to go with UNIX, and the best image processing software at the time was run on a VAX. They had one in the CCIT, in the University computer cluster. What we would do is we would read our images off of a tape, and then we would send them over to CCIT, do the preliminary processing on the VAX, and then send them back, and import them into the software package we were using, which we were busy writing at the time because it didn’t really have much functionality.
Now I’m happy to use other tools to do it, but I’m not afraid to hack and get it to do what I want. But more importantly I know what it should do. Image processing is wonderful because in any other kind of numerical work, if you make a mistake you could be off by a factor of ten to the sixth and you may never know, but in image processing you just look at the screen and you can tell, “Oops.” It’s pretty easy.
|Directory | LARS | LPL Library | LPL WebMail | Webmaster|
Department of Planetary Sciences
Lunar and Planetary Laboratory
1629 E. University Blvd.
Tucson AZ 85721
Copyright © 2008 Arizona Board of Regents