Advanced Vehicle Surveillance System: License Plate Recognition and Tracking.

Abstract: This paper presents a complete framework for a vehicle surveillance system designed for automatic license plate recognition and tracking using deep learning techniques. The proposed system includes a processing pipeline that consists of video processing, object detection, optical character recognition (OCR), and data visualization. For plate detection, we use the YOLOv5 model. We achieve text recognition through a combination of Easy OCR and Tesseract. Additionally, we integrate post-processing algorithms to improve recognition accuracy by enhancing image quality, normalizing text, and validating plate formats. The system includes a web-based interface that allows users to upload surveillance footage, search for detected license plates, and visualize results through an interactive map and analytical dashboards. Experimental results show the system’s performance under real-world conditions with different lighting, camera angles, and vehicle speeds. The proposed approach has strong potential for use in traffic management, law enforcement, and smart city development. Keywords: License plate recognition, deep learning, YOLOv5, optical character recognition, vehicle surveillance, computer vision, image processing.