Determining Road Damage Coordinates Based on the YOLO Algorithm and Microcontroller: England

Aminudin Hamka (Politeknik Negeri Pontianak)
Mariana Syamsudin (Department of Electrical Engineering, Pontianak State Polytechnic)
Medi Yuwono Tharam (Department of Electrical Engineering, Pontianak State Polytechnic)
Hasan (Department of Electrical Engineering, Pontianak State Polytechnic)
Yohannes C. H. Yuwono (Department of Electrical Engineering, Pontianak State Polytechnic)

Infact: Jurnal Sains dan Komputer
Vol. 10 No. 02 (2026): Journal of Science and Computers

Road damage is a critical infrastructure issue affecting transportation safety and efficiency. This study proposes an image- and geolocation-based road damage detection system using an ESP32-S3 camera, GPS module, voltage sensor, and the YOLOv8 object detection algorithm. The camera captures road images at 480×640 pixels in JPEG format, while the GPS module records geographic coordinates and the voltage sensor monitors a 2S Li-Ion battery. Experimental results show reliable system operation. A reference battery voltage of 7.55 V was measured as 7.44 V, yielding a 1.46% error. The GPS module connected to 12 satellites with an HDOP of 0.65. YOLOv8 detected potholes with confidence scores ranging from 0.295 to 0.698, demonstrating the feasibility of integrating IoT sensing and deep learning for automated road damage monitoring.