Eindhoven University of Technology
EDL P16-25 P2: Deep Learning for 3D Reconstruction, Matching &
|Critical vehicle detection in traffic surveillance
Traffic congestion results when the demand for travel exceeds the capacity and when traffic incidents occur. For instance, at critical intersections or in the cities with busy infrastructures, accidents or dangerous goods transportation could occur or be in danger of happening. Recognition of critical vehicles(police, ambulance, fire trucks etc.) will be crucial in that sense and efficient planning and prioritization can be made to go to the place where incident occurred. In that way, the problem will be solved more quickly and increasing the public safety. The vision of a 3D model will be created in the end which integrates both surroundings’ data (mostly static) and traffic data (dynamic data from cameras) of vehicles to manage efficiently the solutions for such calamities mentioned above. TU/e cooperates with the data and platform provider company CycloMedia to create 3D model in this project.
In this research the new dataset will be generated and critical vehicles will be detected and classified based on the requirements. After detecting and classifying the critical vehicles, they will be inserted into the 3D model such that it will give the global situation awareness for critical vehicles moving through the city.