Identifying Parking Lot Occupancy with YOLOv5

Abstract

With the increasing need for efficient parking space management and growing population, the application of artificial intelligence (AI) in occupancy detection has become a topic of significant interest. This paper explores the effectiveness and reliability of the YOLO (You Only Look Once) object detection algorithm in differentiating between occupied and empty parking spots. Moreover, it analyzes the impact of the number of training epochs on the overall accuracy of the AI model. The study utilizes the YOLO algorithm due to its speed and accuracy which makes the training process highly efficient. A custom dataset of 135 images was created and annotated for training purposes. The primary objective of this experiment is to demonstrate the way how AI models can successfully distinguish between occupied and empty parking spaces. By addressing the capabilities of YOLO in occupancy detection, this research aims to contribute to the growing interest in AI applications for efficient parking space management and its implications in tackling real-world challenges.