Automated Kite Detecting Drone Using YOLOv4

• Published in: 2022 6th International Conference on Information Technology, Information Systems and Electrical Engineering (ICITISEE) pp. 600 - 605
• Main Authors: Harfi, Afif Dimas, Rosa, Muhammad Ridho, Romdlony, Muhammad Zakiyullah, Sasmono, Sudarmono
• Format: Conference Proceeding
• Language: English
• Published: IEEE 13.12.2022
• Subjects: Electric Network; Kites; Maintenance engineering; Manuals; Meters; Object detection; Surveillance; UAV; Wires; YOLO
• DOI: 10.1109/ICITISEE57756.2022.10057755

Kites stuck in the electricity network often cause blackouts, especially in West Kalimantan. There were 426 power outages triggered by kite wires in 2018, 219 cases in 2019 and 297 cases in 2020. The method to find the location of kites stuck in the electricity network is still not efficient enough and takes a long time. This happens due to manual observation by deploying some officers to do surveillance along the electric grid. Knowing the position of the stuck kite quickly and precisely using a drone will speed up the maintenance time. The drone is combined with a Raspberry Pi, which has been embedded with a deep learning algorithm, namely YOLO, which will detect kites through a camera connected to the Raspberry Pi. The images captured by the camera will be processed using the YOLO object detection method to detect kites. When a kite is detected, the system will directly store the drone's position (latitude and longitude) and altitude in the database. In this research, we have designed a drone that can detect kites at a distance of 10 to 12 meters at speeds of 1.5 m/s, 3.5 m/s and 5 m/s. The best kite detection accuracy is 88.33% when the drone flies at a speed of 1.5 m/s with a distance of 10 meters.