PHÁT TRIỂN PHƯƠNG PHÁP XÁC ĐỊNHVẬT CẢN CHO DRONE SỬ DỤNG YOLOV8

Abstract

Recently, the use of drones in the Internet of Things (IoTs) is becoming increasingly widespread due to their convenience in performing surveillance tasks. Drones can be integrated with various sensors and utilized in many different fields. However, precise and intelligent drone control is essential to successfully  complete one or multiple tasks. Therefore, this study proposes a new method for drones to determine and avoid obstacle, using a combination of real-time image processing algorithms and deep learning methods based on the structure of YOLOv8 to achieve an accuracy of over 90%.