Description

Environment mapping using computer vision by Ray He

This project attempts to map features in an arbitrary environment using processed video and relative location information recovered from accelerometer data. Furthermore, the project seeks to identify objects based on these features, using combination of Haar-based classifiers, Extended Gaussian Images, and vector representations of segmented sets of features. The eventual goal is to create a reproducible representation of the environment that can be used for vision-based control and provide useful output for an operator overseeing multiple vehicles.