This was part of the ‘SwathSense 2021
’ and ‘MAGIC
’ projects, for which the science plans, implementation, and data collection were designed and coordinated by the NCEO group at King’s College London
, in collaboration with scientists from Europe and USA, while NEODAAS processed the raw airborne data collected into mapped radiances and reflectances.
The NEODAAS team, who has had over a decade of experience processing airborne Earth Observation data, are taking the raw data collected from the new sensor and producing calibrated and georeferenced products which is used by scientists to further develop algorithms.
The Specim AisaFENIX 1K, is hyperspectral imaging spectrometer providing over 600 bands of data, compared to three for a standard camera and 4 – 20 for multispectral satellite sensors. These bands cover wavelengths going from the visible to the short wave infra-red. The wide field of view of the sensor (40˚) and high spatial resolution, with each pixel typically covering 1m on the group depending on flying height) allow for a large area to be surveyed with high spectral and spatial resolutions. Over 4Tb of FENIX 1K data was collected during the flights in 2021, this is the equivalent data volume to over 4,000 satellite images from Sentinel 2.
Example of 23 overlapping georeferenced FENIX 1K flightlines flown in July 2021. High geolocation accuracy
is key to using multiple lines together.
Airborne campaigns are important as they allow data to be collected using sensors which are more advanced than those flown on the current generation of satellites. They also allow flexibility in when data are acquired and experimentation with acquisition parameters (e.g., flight height, direction and instrument settings) which isn’t possible with satellite data. While unmanned aerial vehicles (UAVs), also offer flexibility and high resolution data they are normally limited by the weight they can carry and also the area they cover due to low flight height.
However, processing data from airborne sensors is a complex task, each value from the sensor needs to be matched to a location on the ground and calibrated to provide a consistent ‘real world’ unit (spectral radiance for each wavelength). This involves matching the data collected with the aircraft navigation system, taking into account movements of the aircraft, surface topography and the instrument characteristics. The sensor records a ‘digital number’ indicating the amount of light recorded in each band, calibration coefficients are then applied to convert these to radiance so they can be compared with other instruments.
Will Jay, Airborne Earth Observation Data Analyst at NEODAAS, commented: "The Speim AisaFENIX 1K is a brilliant successor to the standard AisaFENIX,which was used by NCEO to collect data between 2014 - 2019. The below example shows that although the FENIX 1K was recording at 100m higher altitude than the FENIX, it still has much superior spatial resolution with standard FENIX pixels being ~1.5m and FENIX1K pixels being ~0.75m. This also allows flights to be more cost effective, as a larger area can be covered but still achieve the same spatial resolution”.
Side by side comparison of FENIX and FENIX 1K data measured at the same location, Milton Keynes, UK. This false colour composite using wavelengths in the NIR, Red and Green portions of the spectrum highlights productive vegetation in red, with the new FENIX 1K sensor able to show variation within individual trees.
A spokesperson from NCEO commented: “The data from the SwathSense campaign will be used to quantify the extent to which the direction at which remote sensors view the surface impacts the retrieved surface reflectance and land surface temperature - and the extent to which this changes throughout the day with changes in the solar conditions. Hyperspectral airborne sensors are essential to this as they allow us to simulate observations from future satellite missions with different spectral responses.”
NEODAAS look forward to processing data from future NCEO airborne surveys, which alongside the new Specim Asia FENIX will also utilise the Specim AisaIBIS Solar Induced Fluorescence (SIF) sensor. This can provide information on plant photosynthetic activity and the Optech full waveform LiDAR will provide information on the 3-dimentional structure of vegetation, glaciers and volcanos.
’ project investigates how hyperspectral measurements, specifically those required for deriving Land Surface Temperature, are influenced with differing viewing angles across the swath (width of the area imaged).
’, in collaboration with eight partners in Europe, is monitoring methane emissions in the high latitudes of Sweden.