Motor Vehicle Plate Recognition System for Philippine Law Enforcement Checkpoints Using Deep Learning

In most cities of the Philippines, numerous law enforcement checkpoints are stationed as a vital part of securing public order and road safety. As the country’s population grew, we saw a rise in the volume of vehicles on public roads. In areas where a checkpoint is located, the usual flow of traffic becomes congested while officers conduct their standard operating procedures in the verification of vehicles. This study focuses on the development of a motor vehicle plate recognition system for Philippine law enforcement checkpoints. You Only Look Once (YOLO) v4 was used to locate and detect motor vehicle plates through a video feed from an external camera installed in the local machine. OpenCV was used to preprocess the cropped images while Tesseract-OCR was used to extract the motor vehicle plate number as text. The extracted plate number was then used to retrieve the motor vehicle details from a local database. A mean Average Precision (mAP) of 97.67% was achieved after training the model. The experimental results and performance analysis of the study show that the proposed solution is capable of identifying motor vehicles through the recognition of motor vehicle plate numbers.