Deteksi Kerusakan Jalan Menggunakan Pengolahan Citra Deep Learning di Kota Semarang

Abstract

[Title: Road Damage Detection Using Deep Learning Image Processing in Semarang City] Roads are a crucial necessity in community activities. They facilitate transportation access from a particular place to the destination. Furthermore, roads are important as land transportation infrastructure for both people and goods. However, inadequate road conditions can lead to accidents. Tracking road conditions is challenging due to the large number of roads that need to be inspected. This research utilizes remote sensing principles with Deep Learning Artificial Neural Network technology. YOLO (You Only Look Once) is employed for road damage detection. The detection results are enhanced with precise positioning or location using the Global Navigation Satellite System (GNSS), allowing for accurate detection results. This study produces a road damage identification model with an overall accuracy of 88% and a kappa accuracy of 86%, as well as location distribution of damages with positional coordinates accuracy of ±5.6 meters (RMSE).