Unmanned aerial vehicle for hot-spot avoidance with stereo FLIR cameras

In this paper a roadmap is described for future work using stereo thermal cameras to locate and avoid heat sources or hot-spots commonly encountered in disaster remediation tasks using unmanned aerial vehicles (UAV). The proposed stereo forward looking infrared (FLIR) camera system will allow a UAV to locate a body of heat, avoid it to protect its sensitive mechanical and electronic components from excessive heat, and identify the heat source globally to aid remote flight operators. The locations of these identified bodies of heat can be combined with more traditional means of obstacle detection such as LIDAR or a Time of Flight (TOF) camera to allow first responders to deploy a UAV in many different types of emergency events. In addition, first responders can easily operate UAVs semi-autonomously to gather information and act decisively without having to provide obstacle free paths for UAVs in advance. A 3D vector mesh (VM) avoidance algorithm will be used with sensor data from both the stereo FLIR cameras and the TOF camera to achieve both physical and thermal obstacle avoidance. These methods will eventually be combined to achieve a robust and semi-autonomous UAV platform that may provide critical support in typical disaster remediation tasks.