A Dual AI and IoT Approach to Smart Parking: License Plate Recognition and Dynamic Pricing

In this study we introduce an innovative technology of dual artificial intelligence (AI) and Internet of things (IOT) framework for an intelligent parking system website, including automated license plate detection with dynamic pricing tools. The proposed system leverages computer vision algorithms and machine learning models to efficiently monitor and recognize vehicles while dynamically adjusting parking prices based on duration, weather, vehicle types, demand of the slot etc. By incorporating AI-driven components, the solution aims to alleviate traffic congestion, reduce the time drivers spend searching for parking, and enhance revenue management for facility administrators. At its core, the system utilizes YOLOv12, an object detection algorithm, to improve real-time vehicle identification and license plate recognition then perform OCR using TesseractOCR to detect the license plate details. This model effectively addresses challenges related to detecting small, blurry, or large objects through an attentionguided architecture. Additionally, XGBoost, a powerful machine learning algorithm, is employed to refine dynamic pricing strategies by analyzing demand fluctuations and space availability. By seamlessly integrating these advanced technologies, the system enhances overall traffic flow, optimizes parking operations, and creates a more adaptive and automated urban parking infrastructure, ultimately improving efficiency for both users and facility operators.