Have a UROP opening you would like to submit?
Please fill out the form.
Tracking Lots of Particles
IAP and Spring
12: Earth, Atmospheric and Planetary Sciences
Sai Ravela: email@example.com
In this project, we will build a tracker that tracks many particles embedded in a dynamically evolving fluid. There are a few basic requirements: it must work well under varying lighting conditions, changing backgrounds, occlusions, multiple types, and the number of particles, and it must work in real-time. This project will let you start with an existing system and help you become skillful at programming Computer Vision algorithms using OpenCV in C++, and for web/cloud services. The research aspect requires finding fast algorithms for correspondence between many moving features (on regular laptop/phone type computing power). The question is, how many, can you beat the current record? We will try some interesting algorithms, including graph algorithms. The Earth Signals and Systems group has a rich history of developing undergraduate research in the areas of vision, learning, autonomy, and inference for the Earth and planets.
Any course but with experience in C/C++, some knowledge of computer vision, linear algebra, image processing. This could be an ideal thesis project for an undergraduate, and the results will not only enable working in other areas (e.g. tracking problems), but also will directly impact many users who already use an existing (somewhat clunky) system.