This project develops an AI-based system for safe autonomous drone takeoff and landing using computer vision. The system identifies safe landing zones—either marked with visual codes or detected dynamically in unmarked environments, such as during emergencies. Through advanced object detection and scene segmentation, the AI analyzes the surroundings to avoid obstacles like people, vehicles, trees, or power lines, ensuring both operational safety and hardware integrity.
The final integrated software will allow real-time decision-making directly on the drone or via a connected server, combining vision-based perception with intelligent risk assessment.