Abstract
"CubeSat-based Hyperspectral Imagers have become an attractive concept to Earth Observation scientists in recent years, due to the potential of providing cost-effective constellations of satellites with hyperspectral capability. This will be beneficial to applications such as precision agriculture and disaster relief, where increased temporal resolution measurements and rapid response to information request is more critical. However, the challenges of providing science-grade Earth Observation data from CubeSats are great and varied, not only because of the physical size of the instrument, but because of the operational constraints of the CubeSat, such as limited physical pointing stability/accuracy and restricted data downlink budgets.
The objective of this research is to investigate the problem holistically; rather than focussing only on the physical size of the instrument, the operational constraints of the CubeSat platform have been considered since design inception, allowing for the investigation of ways to mitigate these issues within the optical design of the instrument itself. In addition, the cost of the instrument must also be kept within the resources of a typical university CubeSat budget.
To this end, the compact hyperspectral imager prototype CHAFF (CubeSat Hyperspectral Application For Farming) has been designed, developed and constructed with commercial off-the-shelf optics at Surrey Space Centre, in collaboration with the National Physical Laboratory (NPL). As part of the holistic design methodology, CHAFF aims to mitigate the pointing instability of the CubeSat via optically-aided image co-registration, allowing for automatic co-registration and data cube construction on-board the satellite. This in turn will improve the performance of lossless, predictive image compression schemes operating on the data cube, thus mitigating the data downlink bottleneck.
Laboratory and field trials have shown that CHAFF has achieved good optical quality, with a spectral resolution of 3.46 nm at 546 nm. In addition, the image processing chain, including the optically-aided image co-registration, has been observed to improve the motion-induced distortion, by approximately a factor of 12. "