In today’s world, vehicles are the most significant form of transportation. Detection of vehicle number plates is important to identify a specific vehicle. In order to obtain accurate identification of a specific number plate tag of the vehicle, accuracy rate and high precision are critical. Vehicle number plates can be used by law enforcement agencies to track vehicles with traffic cameras that can capture and scan all vehicles in moving traffic. This research proposes a novel system to process digital images of vehicles to capture the license plate number and also identify the type of vehicle using a TensorFlow based algorithm to improve vehicle identification accuracy. The program developed in this study provided a high level of differences for vehicle identification. The program used Python machine learning and TensorFlow deep learning libraries. This approach used the following phases of development: discovery of the vehicle number plate tag, location of the vehicle number plate tag, identification of state of the vehicle number plate tag, reading the number on the vehicle number plate tag, and recognition of the vehicle shape to determine the vehicle type. The proposed method of this work has a correct classification rate of 94.7% in vehicle type identification and 94.4% in license plate number detection, which is much higher than the previous standard methods compared to the proposed method of this work.
Automatic vehicle license plate number detection using machine learning
2020-05-01
Miscellaneous
Electronic Resource
English
DDC: | 629 |