Tour of the UCF Computer Vision Lab

On Monday I went on a tour of the University of Central Florida's (UCF) Computer Vision Lab. Our office is just down the street from UCF and at least half of the people in the office are UCF grads (I'm in the minority - a University of Maryland grad). The purpose of the tour was to talk to some of the students and professors to determine what areas we could collaborate on in the future.

In the photo above, I'm standing next to my division manager, Matt Kraus. We are in front of UCF's entry into the DARPA Urban Challenge (see their page for more photos). Their car is called, interestingly enough, Knight Rider (UCF is home to the Golden Knights in case you aren't aware).

In the photo below, you can see some of the instrumentation they used. On the front of the car are a couple of LIDAR scanners to scan the road, making sure the car stays on track and also is aware of obstacles. The center front of the car has a radar detector that they used to detect oncoming vehicles to help navigate intersections. On top of the car are more LIDAR scanners, a GPS for navigation, and some other gadgets whose purpose I do not recall. They have several rack servers inside to run the instrumentation and control the vehicle.

OK, so those of you who have no idea what the DARPA Urban Challenge was, DARPA offered $2 million to the team whose car could autonomously navigate a 60 mile course in less than 6 hours. The cars were required to obey all traffic rules (no speeding, legally passing, merging into traffic, negotiating intersections, etc.). You read that correctly - autonomously means that the car is driving itself. Carnegie Mellon won the competition in just over 4 hours. UCF, operating on a very small budget, made it 2 hours through the course and then went off track due to a bug in the GPS system.


Actually, the car was on the last part of the tour. The first part of our tour was of the Computer Vision Lab. The students are involved in some very interesting projects, such as an image tracking application that can track a target (person or vehicle). One use for the technology is to locate a person or vehicle in a video feed and track it's movement for surveillance purposes. I'm looking for opportunities to work with them on upcoming projects.

Peace and blessings,
John