Implementasi YOLO untuk Pendeteksi Kemacetan Lalu Lintas secara Real Time

Abstract

Traffic congestion is a serious challenge faced by urban transportation systems worldwide. In an effort to address this issue, the implementation of YOLO (You Only Look Once) object detection technology has gained significant attention. This research builds a system capable of detecting traffic congestion in real-time using the YOLOv8 algorithm to reduce traffic congestion by preventing bottlenecks at road intersections. The study involves several stages, including a literature review to understand the concept of object detection using the YOLOv8 algorithm, dataset collection, labeling, training, evaluation and system implementation based on IoT, utilizing the ESP32-CAM camera to capture real-time traffic conditions. From the testing conducted to evaluate the model's performance using a combination of 100 epochs with a learning rate of 0.01, the highest mAP value obtained was 0.849 for all classes, and the F1 confidence score was 0.713. This research demonstrates that a system using the YOLOv8 model is capable of detecting vehicle objects and analyzing conditions at road intersections in real-time..