DETEKSI PELANGGARAN LALU LINTAS DI JALAN TOL MENGGUNAKAN FRAMEWORK YOLO DAN KALMAN FILTER
Fathan Qoriba Hanif, Rudi Heriansyah, Zaid Romegar Mair · Jurnal Publikasi Teknik Informatika · 2025
Traffic violations on toll roads remain a serious challenge in developing a safe and orderly transportation system. This study aims to develop a traffic violation detection system based on CCTV recordings by integrating the You Only Look Once (YOLO) algorithm for object detection and the Kalman Filter for vehicle tracking. The data used were obtained from the Kapal Betung toll road KM 362, South Sumatra. The system was tested to identify speeding violations and improper lane usage by trucks. The results showed that YOLOv11 achieved a detection accuracy of 89.4% mAP, while the Kalman Filter provided 94.2% tracking accuracy and 96.1% average speed estimation accuracy. Lane violation detection by trucks reached a verification level of 88.5%. This system demonstrates strong potential for real-time traffic surveillance and serves as a foundation for the development of smart transportation technologies in Indonesia. Nevertheless, challenges remain in extreme visual conditions, indicating the need for future systems to incorporate multi-sensor approaches.