IoT-Based Automatic Transaction of Smart Parking Using Embedded System and Machine Vision

The efficient parking system is still a big challenge due to the growth in the number of vehicles while the parking areas are limited. This work aims to develop a seamless smart parking transaction system by leveraging artificial intelligence and IoT technologies. The proposed system uses object detection and image processing to make a new transaction system based on user vehicle licenses. The system is designed into a five-layer architecture that consists of several functions and modules, i.e., the license plate detection and recognition system, hardware embedded system using microcontroller, communication system, web application system, database, and backend Application Programming Interface (API). IoT-based backend technology integrates functions and facilitates user interactions, such as user account activation and checkout from parking areas. The proposed system uses YOLOv5 for license plate detection and Tesseract OCR for character recognition. The system efficiently detects incoming and outgoing vehicles by capturing their license plate numbers. Upon license plate detection, the system automatically charges the corresponding user's account, streamlining and reducing the parking transaction process. The average processing time for the check- in phase is 10.4 seconds, starting from the camera performing license plate recognition until the data is stored in the blockchain. On the other hand, the average processing time for the checkout phase is 11.1 seconds. Meanwhile, YOLOv5 has 99.63% accuracy in detecting Indonesian plate numbers.