For this project, we created a software package to run on the Hack Pittsburgh Autonomous Car. The goal was to classify identified and track static or dynamic objects around the vehicle as either human or non-human in real-time. These classifications will update online as more information becomes available, objects overlap and merge with each other, and objects move in and out of the frame.
The goal of this project was to build an online zero-finding algorithm within the Simulink architecture. I created a novel online version of Muller’s zero finding method and a progressive-guess system which makes an initial prediction at a time sufficiently far away, then tweaks the trajectory as it approaches zero.
My final project for Computer vision was a program designed to turn any cellphone camera into a 3D scanner. I was able to take individual pictures from an uncalibrated monocular camera at arbitrary positions around an object and construct a 3D representation of it.
