IoT Based Real-Time Automatic Number Plate Detection Using OpenCV

In recent times, there has been a high increase in vehicles plying on the road, due to the high increase in population as well as other geographical constraints. Consequently, an accurate vehicle detection system for traffic control is required. With the advancement and development of powerful and low-cost single-board computers, developing a fully automated system is becoming much easier than ever. An embedded system capable of executing deep learning models for object detection has been implemented in this paper. The main aim of this paper is to develop an Automatic Number Plate Recognition (ANPR) system that recognizes license plates using a deep learning object detection model and optical character recognition (OCR). A Raspberry Pi with a camera is used in the setup to automatically recognize number plates. Multiple techniques have been used to deploy on the Raspberry Pi which includes the implementation of state-of-the-art object detection models such as Single Shot MultiBox Detector (SSD) Mobilenet, libraries such as OpenCV, making it a standalone system that does not rely on another device to do the processing. The data is sent over the internet using Google Cloud APIs. This makes the system a low-cost, fully automated system that is portable and can be easily deployed at any remote location.