CycleCV

This project contributes to the idea of using computer vision to analyze urban traffic and improve urban transportation planning, with particular focus on cycling. The outcome of the project is a data annotation platform aimed at reconstructing a dataset of traffic participants trajectories at one of the intersections in Copenhagen.

This project builds on top of my research project at ITU. During my research project, I developed an algorithm for reconstructing urban traffic trajectories in a dataset based on a 10-hour video footage.

The annotation platform was built to validate the algorithm and create labeled data to further improve it. The plaform provides an interface to label trajectories and make alternative trajectories.

Result

• Reduced time effort to annotate a large dataset
• Enabled further research and analysis of urban traffic using CV

Scope

Technologies