There is a growing need for efficient transportation operations and communication, especially for urban emergency evacuation scenarios such as natural disasters and severe weather conditions. The goal of this SuperUROP project is to build a convolutional neural network using the Python-based tensorflow framework for traffic-related objects using publicly-available semi-unstructured camera data from MassDOT and CalTrans. The model will be used to classify over a million images for future transportation modeling research about disruptions in critical transportation facilities due to natural events.

“I am participating in SuperUROP for the opportunity to apply machine learning techniques I have learned in 6.036 (Introduction to Machine Learning) to real traffic data. I hope to improve my programming skills and machine learning knowledge and contribute to interesting research about urban transportation.”