Vision Transformer based Intelligent Parking System for Smart Cities

In order to address the issues brought about by urbanization, population growth, and the demands of sustainable cultural development, the idea of "smart cities" was developed. One crucial element of smart cities that helps manage and lessen traffic is smart parking systems. Parking spots are a limited resource, and in crowded areas, finding free spots can be difficult. We provide a real-time free parking slot detection system that uses computer vision and machine learning techniques to locate and identify free parking spaces in real-time, thereby resolving the current issue. Our system uses a camera network deployed in the parking area to capture images of the parking slots. These images are processed using deep learning combined with vision transformer and transfer learning algorithms to detect the presence of vehicles in the slots. By analyzing the occupancy status of each slot, our system can identify and locate free parking slots. We evaluate the accomplishment of the model we have offered on a real-world dataset collected from a busy parking area. The findings demonstrate that the offered model outperformed the CNN model with a test accuracy of 95%, while CNN achieved an accuracy of 90%.