Error assessment and mitigation for hyper-temporal UAV-borne LiDAR surveys of forest inventory.

Remotely sensed LiDAR data has become an important tool in the management of modern forest inventories. Monitoring the high frequency changes within forests with this data has been restricted by the cost and intermittent nature of LiDAR surveys. The use of Unmanned Aerial Vehicles (UAVs) as a remote sensing platform is a rapidly developing field and is capable of allowing highly dynamic environmental changes to be monitored. As such recent studies presented in the literature highlight the potential of UAV systems for forest monitoring. This study further investigates the potential of UAVs by examining the achievable accuracy of a newly developed UAV-borne LiDAR system in comparison to a traditional full scale system. The major contributions to the error budget of a UAV-borne LiDAR system are constrained through the use of a novel UAV specific processing workflow. Central to this workflow is the fusion of observations from a low cost Inertial Measurement Unit, a GPS receiver and a high definition video camera with a Sigma-Point Kalman Smoother allowing for highly accurate estimates of orientation. We found that using this workflow and under certain flying conditions accuracies similar to a modern full-scale system are achievable from this low-cost platform.