Real-Time Parking Space Detection and Management with Artificial Intelligence and Deep Learning System

Shweta Shukla,Rishabh Gupta,Sarthik Garg,Samarpan Harit,Rijwan Khan

DOI: https://doi.org/10.1007/978-3-030-86749-2_7

2022-01-01

Abstract:A rapid rise in the number of cars running on roads has increased the need for parking spaces due to which many difficulties are faced by drivers to find a parking space in car parking areas in big cities. Utilizing the capabilities of computer vision, the authors propose an automated parking space detection and management system. This chapter presents an approach for classifying parking lots either filled or empty based on the Convolutional Neural Network (CNN) running on the administrator’s or owner’s system. Our application provides an efficient and robust way of occupancy detection in various light conditions, weather conditions, presence of shadows, and partial occlusions. The authors used a publicly available CNRPark dataset that contains nearly 9000 images captured from various cameras in different conditions and with obstacles. Along with this, the LSTM model is trained based on the classification data provided by the CNN model collected over a long period to predict the future availability of parking lots based on the specific day of the week, festival, and national holiday. Additionally, a deep learning model is used as an alarming system to detect unwanted events occurring in the parking space. This security system which detects oddity helps to improve the reliability of the system and cameras. No manual monitoring is required due to this anomaly detection system which takes the same input as the slot detection model. All these features make our system more efficient, reliable, and cheaper than traditional methods of parking space management systems such as barriers, sensors, etc. This system is served to the user through a web app and Android app from where the user can gather useful information.