top of page

I am the lead for the Traffic Camera Dangerous Driver Detection (TCD3) project, and conceptualized and implemented the project from its inception. The project is designed with one goal in mind, end drunk and distracted driving. After my uncle was hit by a drunk driver, I decided to work to determine some way to solve the age-old problem of drunk driving. I decided to use traffic cameras, as they are already installed in many communities to detect these drivers. My software works by using various Computer Vision methods to detect cars, and determine their position. The software then intelligently builds a model of a normal driving, taking into account weather, traffic patterns, and the location of the traffic camera. The software then detects drunk drivers in real time, and sends a picture of their car, and location, to law enforcement for further action. This software has the ability to stop a critical problem, leveraging the existing sensor network.

 

I approached the University of Dayton Research Institute, as they had worked with the Dayton Police Department with Traffic Cameras in the past. I decided to continue working on their Computer Vision applications, including mine, and work to roll the software out for real use. In addition to the TCD3 project, I worked to use my experience to benefit their other projects. UDRI had a contract with Honda to develop high speed thermal cameras to intelligently, and in real time, cool down metals as they were cast to reduce hot spots that cause part rejects, as they are brittle.

 

At UDRI, I was able to propose, and work on, my own project, and enjoyed the opportunity to be able to work on my own research project. UDRI allowed me to not only learn technical skills, but learn how research labs work. This experience was critical in gaining critical experience in collaborating with experts in the field. 

 

I continue to work with UDRI to roll out my system in Dayton, and am planning to launch a beta version in Early 2016. Working at UDRI was also a valuable networking experience, as I was able to work with experts in the field of Computer Vision and Machine Learning. 

 

A screenshot showing TCD3 detecting a car using traffic camera footage

A video explaining TCD3

Complete information about this project is available at drunkdriverdetection.com.

bottom of page