Revolutionizing Road Safety: Deep Learning in Vehicle Detection on Rs Fatmawati Road

Muhammad Syafiq Reynara (Faculty of Management and Business, Institut Transportasi dan Logistik Trisakti, Jakarta, Indonesia)
Oktaria Dwi Yanti (Faculty of Management and Business, Institut Transportasi dan Logistik Trisakti, Jakarta, Indonesia)
Abdullah Ade Suryobuwono (Faculty of Management and Business, Institut Transportasi dan Logistik Trisakti, Jakarta, Indonesia)
Prima Widiyanto (Faculty of Management and Business, Institut Transportasi dan Logistik Trisakti, Jakarta, Indonesia)

Jurnal Ilmu Multidisplin
Vol. 3 No. 4 (2025): Jurnal Ilmu Multidisplin (Januari-Maret 2025)

This study evaluates the YOLOv8 deep learning model for vehicle detection on Jalan RS Fatmawati, Jakarta, using publicly available CCTV footage. Extensive pre-processing and data augmentation enhance model robustness, with performance assessed through metrics like mAP, precision, recall, and F1-score. Results indicate YOLOv8's high accuracy and reliability in real-time vehicle detection across various weather and lighting conditions, offering significant implications for urban planning and traffic management. The research also compares YOLOv8 with previous models (YOLOv4 and YOLOv5), revealing superior performance under specific conditions but challenges in extreme environments. These findings underline the model's practical utility and identify areas for future research.