This project is developing techniques for high-resolution true video orthoimage generation from Unmanned Aerial Vehicle (UAV) video data in forest areas. There are many technical challenges in generating true video orthoimages. To address these challenges, the project is establishing a rigorously photogrammetric self-calibration model to: (a) remove the relief displacement caused by trees; (b) compensate for the occluded objects, such as houses and roads, which are usually important indicators of life existence during firefighting; and (c) increase the accuracy of UAV sensor's exterior orientation parameters and lens distortion determination, which are solved simultaneously.
The objectives of the project are: (1) to provide highly accurate and reliable forest fire surveillance data for forest fire experts, analysts, decision-makers and firefighters in order to increase the efficiency of monitoring forest fires to improve forest fire management and mitigate forest fire disasters; (2) to increase and extend the applicability and usability of the UAV video images in fast response to time-critical events and natural disasters.
The broader impacts of the project include: (1) the proposed methods are significantly beneficial to the future U.S. UAV application strategy in areas such as surveillance, real-time response to time-critical events, environmental monitoring, etc; (2) the proposed technique can be directly used by U.S. industry, thus the project significantly increases and extends the usability of commercial UAVs.
