Improve the data quality for effective traffic signal control by utilising the video and sensor data.

My current research in the University of Huddersfield is targeted at traffic analysis and control using Computer Vision, relying on data from CCTV cameras. Computer vision is a field of AI that looks at automatically analyse images or videos. Focusing on traffic, computer vision plays a major role in tracking vehicles, detecting accidents and congestion, and monitoring the level of service of a network.

Transport For Greater Manchester (TfGM) is responsible for managing various transport services across Greater Manchester. This requires collating data from various sources such as sensors, CCTV cameras etc.

With this opportunity, I would like to increase my understanding on Traffic Engineering application which make use of the data sourced from CCTV camera. In particular, the secondment would focus on 2 main aspects:

Exploring how CCTV camers are installed and deployed at junctions, and how their performance are assessed;

Developing a better understanding of how CCTV cameras’ data are accessed and stored.

The second point above is of particular importance, as it allows to understand whether computer vision analysis needs to be performed in real-time, or can be performed offline, with significant impact on performance and technologies to be considered.

Working at TfGM would allow me to get insights into the use of CCTVs and corresponding data, and discuss with experts about the current way in which computer vision is currently exploited, and how it can be improved to better suit the needs of traffic engineers.