IoT based Parking Management System using Video Processing for Open Space Parking Area

Abstract

The flawed parking management system may waste time and cost searching parking spaces. It may also cause traffic congestion and carbon emission. Finding a parking space in most metropolitan areas, especially during rush hours, is difficult for drivers. A difficulty arises from not knowing where the available spaces may be when traffic congestion may occur. The proposed system avoids poor parking management through object recognition technology that detects parking spaces of modern busy parking lots. Traffic congestion reduces by notifying the driver exact vacant parking spaces location. So the users can park their car in available space directly without looking around for space. The main objective of this project is to design IoT based parking system using TensorFlow based on object detection technology. Follow by the integration of the developed IoT parking system for normal road user’s. The project used Anaconda prompt to run the TensorFlow main object detection code. Telegram app act as IoT platform for users that can notify about the available or non-available parking. For the demo purpose, the project uses a video feed as the input for object detection. As a result, the parking status data keeps counting on the terminal for object detections in Anaconda prompt. At the same time, Telegram sends the notification to the user. In conclusion, the system proposes a practical methodology for managing the parking system using a camera placed for a lamp-post view. The system’s accuracy depends on the training model used for object training and the number of samples provided for training.