The project aims to develop an automatic sound detection and classification system using machine learning applied to the management of land transport infrastructures. In ROSOMACHINE, we propose an applied research and development of sound classification on roads in order to expand the current possibilities of road and traffic management. Many of the events that road management wants to know in the most reliable and automatic way possible can be addressed from machine learning applied to acoustic data.
Therefore, if we are able to identify and classify events by their sound, we could monitor in an innovative way, controlling them:
-Traffic safety by detecting, classifying and quantifying the sounds associated with impacts between vehicles, exits from the road, sudden braking, etc.
-Traffic flow control thanks to the sound associated with different traffic speeds, idling engines in traffic jams, etc.
-Classification and quantification of the different types of vehicles in transit, differentiating and counting motorbikes, passenger cars, heavy vehicles, etc.