Autonomous Quadcopter Image Processing for Simulated Search and Rescue Flights

Abstract

In Search and Rescue (SAR) operations, a designated search area is explored with aircraft and helicopters. After target identification, the flying vehicle continues landing and rescue procedures. This research uses a quadcopter to replicate a SAR flight simulation. Autonomous quadcopter operation includes takeoff and navigation between waypoints determined by Mission Planner software. On the way to the second waypoint, a camera-based image processing system scans the ground surface. If the marker is detected by the image processing system, the Raspberry Pi program will instruct control commands to the Pixhawk flight controller to ensure the quadcopter lands directly on the recognized marker. In the case when the quadcopter reaches the second waypoint but the system fails to identify the marker, the Mission Planner commands the quadcopter to autonomously return to the starting point and land automatically at the take-off location. An interesting aspect of this research concerns the application of a low-cost image processing system to ensure the quadcopter flies at a constant flight altitude above the ground surface, so that the quadcopter can perform simulated SAR flight missions and accurately identify landmarks. Research parameters include marker diameter, flight altitude, and quadcopter speed. The results show successful marker detection at a flying altitude of up to 3 meters above the ground and reaching a top speed of 3 m/s at a flying altitude of 2 meters.