Daeun Song, Bo Yoon Choi, Soyoung Chung, Byong-Gul Lee, Journal of Internet Computing and Services, Vol. 27, No. 1, pp. 99-112, Feb. 2026
10.7472/jksii.2026.27.1.99, Full Text:  HTML
Keywords: Autonomous Flight Drone, fog detection, Lightweight Real-time Perception, Urban Navigation Safety

Abstract

Although autonomous flying drone technology has advanced rapidly, extreme weather conditions such as dense fog and heavy rain continue to pose significant limitations by degrading sensor performance and reducing navigational visibility. This study proposes a real-time dense fog detection framework to enable stable autonomous flight in foggy environments by utilizing sequential data from both camera and LiDAR sensors. First, dense fog is detected in real time by calculating the Dark Channel values from camera images. Subsequently, LiDAR point cloud data are analyzed to estimate the range of point loss, allowing the system to identify dense fog regions along the flight path. By processing the drone’s camera and LiDAR data separately and conditionally, the proposed method reduces the complexity of sensor fusion and minimizes computational load, enables real-time data processing. Experiments conducted in an urban environment demonstrate that the framework enhances the safety and efficiency of autonomous drones by ensuring real-time performance, robust detection, and computational efficiency.

Statistics
Cite this article