Drone-Based Path Localization for Surface Mobile Robot with 2D Camera

In water surface search-and-rescue operations, where lives are at extreme risk, a life-saving boat with only a surface-level view can be ineffective, slow, and may miss individuals in distress. By incorporating a drone equipped with a camera, the search area coverage can be significantly increased, and the entire scene can be observed, enhancing the efficiency and effectiveness of the rescue efforts. This paper presents a drone-based path localization system for guiding a surface mobile robot using a 2D camera. A drone equipped with an RGB camera is employed to detect and localize goal points marked by ArUco markers on a surface grid. The system calculates the real-world coordinates of the goal points relative to the mobile robot's position from the images captured by the drone's camera. The experiments were conducted at different drone heights and camera angles to assess the system's accuracy and robustness in different environment. The results demonstrate that the drone can accurately predict the real-world coordinates of points captured by the 2D camera. The system has 2% to 5% errors across all drone altitudes and camera angles. The system's performance suggests its potential application in search and rescue operations, where accurate and reliable navigation is critical.