Deteksi Jalan Berlubang Menggunakan Algoritma Yolov5

Abstract

The city of Sorong, as one of the largest cities in the Southwest Papua region, is facing serious problems due to potholes in various areas. This problem causes the risk of accidents and vehicle damage, disrupts the mobility of city residents, and hinders sustainable infrastructure development. Therefore, we need a system that is efficient and accurate in detecting potholes quickly. In this study, researchers used the "You Only Look Once" (YOLO v5) method to detect potholes in Sorong City. YOLO v5 is a real-time object detection algorithm that has been proven to have high speed and accuracy in recognizing objects. This approach allows for instant detection of potholes as vehicles pass by, so authorities can quickly take the necessary remedial action. This study aims to implement pothole detection technology using YOLO v5, with the hope of increasing efficiency and accuracy in overcoming potholes in Sorong City and improving traffic safety by detecting potholes in real-time. The results of this study obtained the highest training precision, namely 96.6% and the accuracy value of the pothole detection system test.