AUTOMATIC TRACKING
OF A VEHICLE
We’ve conducted a study to assess the feasibility and measure the outcomes of automatically tracking a vehicle moving at varying speeds (up to 300Km/h) with static cameras and visual analysis software containing an artificial intelligence component. The cameras have been installed in key positions to be able to monitor every section of the circuit.
It is anticipated that the vehicle will be re-identified by the cameras based on its position on the track (for example it is expected that the vehicle is first identified by camera 1, then by camera 2 and so on).
For a fault-tolerant system, however, sensor fusion techniques (GPS + Camera) was used to verify the vehicle’s position on track at any time.
The software have been developed along a period of 18 months and refined with “real conditions” data and live tests conducted in various lighting and weather conditions. The results obtained and shown in the video below, display high accuracy in vehicle recognition without considerable data loss.
