Towards Smarter Parking:An End-to-End System for Real-Time Space Detection and Path Finding

Parking congestion is still a key problem in cities, resulting in excessive fuel use, wastage of time, and environmental issues. This paper offers an End-to-End Smart Parking Assistance System combining image segmentation, graph neural networks (GNNs), and swarm intelligence to identify vacant parking lots and navigate the vehicle accordingly effectively. The methodology starts with image segmentation, where parking lots are correctly detected from aerial and ground images. Processed spatial data is then modelled by a Graph Neural Network (GNN) to forecast optimal space availability in parking lots. A Swarm Intelligence-based pathfinding algorithm lastly dynamically calculates the optimal path to a parking place in real time, reducing transit time and traffic. The methodology outperforms traditional parking systems because it provides real-time adaptive and optimized parking assistance. Experimental outcomes illustrate better space detection accuracy, more efficient navigation, and better overall parking management than conventional techniques.