Improving the Spatial and Spectral Calibration of Remote Sensing Imagery from Unmanned Aircraft Systems

Remote sensing (RS) using small unmanned aircraft systems (UAS) is a common strategy for collecting site-specific measurements in precision agriculture. UAS carry a wide range of imaging sensors and can rapidly collect spatial and spectral data at spatiotemporal resolutions that were previously challenging to achieve. UAS deployment, image acquisition, and photogrammetry are largely automated process. While the technique has gained a large footprint in the research domain, technical barriers associated with spatial and spectral calibration of remote sensing imagery limits application in production agriculture and continues to present bottlenecks to field research. Calibrating RS imagery is far less automated, which presents scalability issues when transitioning from research to production and when comparing independent research efforts. Limited guidance is available for researchers on how to collect data that is accurate – specifically, using low-cost systems. Likewise, limited knowledge is available to producers on how to use UAS-based RS imagery to make actionable decisions. To address these gaps, our proposal seeks to investigate automated techniques for calibrating RS imagery. We will measure the spatial accuracy of RS products when using active and passive ground control targets and from camera systems with varying georeferencing accuracy. We will address spectral accuracy by augmenting existing RS image acquisition with reference measurements from onboard spectrometers, terrestrial pyranometers, and ground reflectance targets. These combined efforts aim to improve the accuracy of RS derived measurements so that models developed by researchers can be employed by producers at scale.