Optimalisasi Sistem Deteksi Parkir Menggunakan YOLOv7 dan CLAHE

Abstract

The significant increase in the number of four-wheeled vehicles in Indonesia, particularly in urban areas, presents challenges in the efficient management of parking spaces. One potential solution is the implementation of an automated parking detection system based on digital image processing and artificial intelligence. This study aims to develop a prototype system for detecting vacant parking spots using the YOLOv7 algorithm, optimized with the Contrast Limited Adaptive Histogram Equalization (CLAHE) method. The image dataset was captured from a single-level parking prototype under various lighting conditions. CLAHE preprocessing was applied to enhance the visibility of essential image features, particularly in low-light scenarios. The training results showed that the use of CLAHE improved detection accuracy compared to YOLOv7 without optimization. Evaluation metrics included precision, recall, F1-score, and mean Average Precision (mAP). The developed system demonstrates significant potential in enhancing the efficiency of parking space usage, reducing search time, and lowering vehicle emissions. This research contributes to the advancement of efficient and environmentally friendly smart parking technologies