Optimal UAV movement control for farming area scanning using hyperspectral pushbroom sensors

Abstract

There is a growing interest in setting hyperspectral cameras on UAVs for periodically scanning farming areas with the goal of obtaining more information about the crops and soil state and optimize the productivity of the agricultural field. Hyperspectral pushbroom scanners offer a good trade off between spatial and spectral resolution for such purpose. However, this kind of hyperspectral cameras also introduces technical difficulties mainly related to the lack of intraframe spatial information, turning the registration of subsequent frames a hard task which is typically carried out off-board, once the flight mission has ended, preventing the posibility of obtaining real-time results. This work tackles this problem by controlling the aircraft flight movement in such a way that meaningful hyperspectral images can be directly obtained by just stitching the subsequent captured frames together. The main goal is to offer the possibility of analysing the acquired data in real time, avoiding the necessity of carrying out any registration process. The developed UAV movement control has been tested on real flight campaigns over a vineyard in the island of Gran Canaria. The obtained results verify the goodness of our proposal.