This summer, I'm working with NASA's Jet Propulsion Lab. We're extending the computer vision capabilities of the Mars Rovers. The main idea is that the rover will be able to determine what information is actually important in an image, tag the important segments with metadata, and report only the relevant/important images and information. For example, if we're looking at boulders, having an image that also includes part of the sky is wasting a lot of bandwidth. I thought WiFi was slow, but all the way from Mars just plain sucks. All our real code will be in C and a good portion of it will be systems level. Simulations and prototyping will be in Matlab and a couple in-house tools.
My previous project (which is now complete) was with NIST and the Department of Defense. It was an automated tool for performance evaluation of video surveillance algorithms, for all sorts of video including broadcast news, UAV's, fixed cameras, etc. It involved a lot of machine learning, pattern recognition, etc.
I bring these up because this is what I like. I really enjoy these kinds of projects and I tend to catch on to things easily. So my new question is: if this is what I want to do, am I stuck in academics for the rest of my life? Should I get my doctorate and just do research at a university? What about the private sector? And I don't know... something about working for our government kind of scares me.