Automation enforcement on priority lanes

Abstract

Constantly, traffic laws are being violated, this is mainly because enforcement is not monitored 24/7 and hence it makes it easier for drivers to infringe the law. However, this problem may be mitigated by using Automated Enforcement Systems (AESs), as this replaces the traditional method of how enforcement on the roads is carried out. Currently, in Malta, AESs systems are implemented for speed cameras and red-light running. The scope of this project is to create an AESs to tackle the enforcement on priority lanes by detecting vehicles that does not fall in the following categories: Motorcycles, LPG cars, Electric vehicles, Route buses, Passenger transport vehicle, Taxis, Pedal cycles, Vehicles on priority duty (ambulances, police cars etc.) and vehicles carrying more than three persons in a vehicle. An overview of the system would compose of first detecting a vehicle on the priority lane. Tracking of this vehicle is then triggered to follow the car in the video scene. Automatic number plate recognition (ANPR) is then used to extract the number plate to be able to identify if the driver is infringing the law. Finally, a short clip of the scene is exported to be send to enforcement units for further verification. Upon testing the whole system for different scenarios, a vehicle detection accuracy rate of 90% was achieved. For the recognition of characters in the ANPR function, an 84% accuracy was noted. Improvements and limitations to justify the results obtained for this system, are then further discussed in the conclusion chapter.