Does getting into a mall, waiting to get a parking spot, queuing up to pay the parking fee wear you out? These students from Chennai's Great Learning might have a solution. Five students of the AI, ML PGP course developed an integrated and automated vehicle parking management system using Computer Vision techniques to optimise the available parking spaces. The computer vision-enabled system uses cameras at entry, exit and inside parking spaces to automate the entire parking process without manual intervention from entry to payments and exit without stopping the vehicle or installing any additional equipment in the car.
Developed by Great Learning’s students Faisal Khan, Kiran Mogeraya, Prasanna Siva, Praful Mohan Agadi, Shubham Mandal, the research has also been featured in the 11th IEEE Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON) on December 25, 2020. The research is also a part of the Capstone project, which allows Great Learning students to apply their learning to real industry projects and add it to their portfolio for potential employers. "We wanted to build something that can be utilised and easily adopted into the market," says Prasanna.
How does it work?
Wherever a parking system and vehicles are involved, this system is pretty useful. "If someone is going inside a mall then you have to queue up, currently there are people to assist in larger malls, it's a long process. To avoid all that, the system has been made automated. As soon as the car enters there's a camera that will detect the license plate number, there will be a database repository will be maintained. At a large complex or society, with this every time a car enters the same place, it will detect if the car is a residential one or outsiders. At places like malls, there is a computer vision-based system that we developed which will detect the license plate, it will then help navigate the car to a free parking spot, at the exit again the automated system will calculate the hours and the parking fee. There is no manual intervention and makes it a lot faster," explains Prasanna and Praful.
The automated system generates the information about free parking slots by identifying the parked cars as objects and then deducing the free slots using the total available slots in that area. "At the exit, the license plate information collected earlier is used to detect the parking duration and to facilitate automated fee collection. The system also offers real-time vehicle surveillance for high reliability and vehicle security and ensures hassle-free parking management and customer satisfaction," adds Faisal.
The students claim that the model has demonstrated 100 per cent accuracy in license plate detection and free parking slot identification and was trained on more than 300 images and tested on over 20 car images covering various data-capture scenarios.
