Vision-based Navigation of Unmanned Aerial Vehicles

Unmanned Aerial Vehicles (UAVs) are increasingly being used for a wide range of civilian and industrial applications. While GPS-based navigation systems are commonly employed, they suffer from limitations in environments where GPS signals are weak or unavailable, such as indoors or in dense urban areas. This paper presents a vision-based navigation framework that leverages visual data for real-time localization and path planning, providing an alternative to GPS navigation. The proposed method uses a camera system to capture visual features, building a visual memory of the environment during a guided flight. This memory is then used for localization and autonomous navigation in real-time. Experiments conducted with an indoor UAV equipped with a fish-eye camera demonstrate the effectiveness of the approach, showing reliable navigation in complex environments without the need for GPS. The results highlight the potential of vision-based systems for enhancing the autonomy and versatility of UAVs, particularly in GPS-denied scenarios.