A Compact, High Resolution
Hyperspectral Imager for
Remote Sensing of Soil
Moisture

Abstract

Measurement of soil moisture content is a key challenge across a variety of fields, ranging from civil engineering through to defence and agriculture. While dedicated satellite platforms like SMAP and SMOS provide high spatial coverage, their low spatial resolution limits their application to larger regional studies. The advent of compact, high lift capacity UAVs has enabled small scale surveys of specific farmland cites.

This thesis presents work on the development of a compact, high spatial and spectral resolution hyperspectral imager, designed for remote measurement of soil moisture content. The optical design of the system incorporates a bespoke freeform blazed diffraction grating, providing higher optical performance at a similar aperture to conventional Offner-Chrisp designs.

The key challenges of UAV-borne hyperspectral imaging relate to using only solar illumination, with both intermittent cloud cover and atmospheric water absorption creating challenges in obtaining accurate reflectance measurements. A hardware based calibration channel for mitigating cloud cover effects is introduced, along with a comparison of methods for recovering soil moisture content from reflectance data under varying illumination conditions. The data processing pipeline required to process the raw pushbroom data into georectified images is also discussed.

Finally, preliminary work on applying soil moisture techniques to leaf imaging are presented.