Commercial hyperspectral is entering the market with the first two low-cost satellite systems having already been deployed. The main challenge is the same as for optical multispectral or SAR constellations – to build cost-effective systems that can provide high-revisit data for building change detection analytics. More specifically for hyperspectral systems, the challenge lies in finding the sweet spot between spatial and spectral resolution, as well as the number of bands and the right spectrum bands to create new commercial applications.
We have identified more than 10 hyperspectral companies that have announced plans to develop commercial constellations over the next five years. However, two key historical barriers have to be overcome for Hyperspectral Imaging to break into the mainstream commercial market: the cost of data and processing complexity. “What is promising is that these are the same barriers that are successfully being overcome by SAR companies such as Iceye”, said Alexis Conte, Editor of the Earth Observation Data and Services Market research report and Senior Consultant at Euroconsult.
The cost barrier
In-house manufacturing, access to space, and competition are favorable exogen factors driving costs down. However, to be disruptive and low-cost, systems are requiring trade-offs on the performance level.
Hyperspectral technology has been mainly used for decades by airplanes and a dozen or so civilian government satellites. Hyperspectral technology implies collecting data across a large number of spectral bands (generally more than 100) and then delineating them into many fine spectral bands (< 10nm) contiguously. These imaging spectrometers are capable of acquiring a continuous spectrum at a given wavelength range generally into the VNIR-SWIR domain and possibly to the TIR domain (800-12,000nm). However, new hyperspectral systems such as Orbital Sidekick (US), HySpecIQ (US), HyperSat (US), Zhuhai Orbital OHS (China), Satellogic (Argentina), and SatRevolution-II type (Europe), are low-cost systems capable of selecting only ~30 bands over a specific part of the spectrum, generally into the Visible and Near-Infrared range (VNIR).
The hyperspectral ability to switch bands differs from multispectral fixed-band systems. As most applications require a small number of specific bands, and as these bands are not the same for all applications, this ability to switch adds value compared to the most performant multispectral systems (from generally four to exceptionally 28 fixed-bands for WorldView-3). Being able to focus on only 30 bands across the VNIR range (400-950nm) is likely the key to reducing cost of traditional hyperspectral systems. Then, efforts can be made to improve spatial and spectral resolution performances to better identify objects’ characteristics and meeting a wide range of end-user requirements.
Amongst the two commercial hyperspectral systems having already been deployed, Satellogic started with 25m hyperspectral spatial resolution as a secondary payload from its Ñusat series. The system has 18 satellites currently in orbit (as of December 2020). Satellogic plans to incrementally increase this resolution through time with a new generation of satellites to be launched. Regarding Zhuhai Orbital, 8 satellites out of the 10 planned of the OHS series are already operational. The 67kg platform provides a 10m spatial resolution acquiring in 32 bands simultaneously at 2.5nm spectral resolution across a 400-1,000nm range. The constellation covers the entire globe every two days with a swath of 150km along 2,500km per image strip.
The processing complexity barrier
These high-resolution, low-cost systems are opening up new markets by focusing on specific applications and delivering turnkey services, fostering end-user adoption, and increasing the customer base. This is the second historical barrier to be overcome. With embedded service offerings, there is no need to be a technical expert in image processing to extract the spectral signature of interest, bringing this to the mainstream market. As hyperspectral added value is a challenge to extract, Off-The-Shelf services are the ultimate way to propose a cost-effective solution to multiple users as long as it specializes in specific applications. For instance, hyperspectral company Orbital Sidekick is paving the way for pipeline integrity monitoring via Satellite, while most of the current applications in this domain are relying on aerial, drone, and field surveys that are costly and hazardous to perform. The firm awarded a contract to demonstrate how satellite-based value can detect and prevent leakage over two oil basins in the US on a daily basis monitoring. This will be performed by information gathered from aerial and third-party satellites while the company plans to build out its full satellite constellation over the next 18 months.
Low-cost constellations and integrated services pave the way for satellite-based hyperspectral
Greater opportunities will come from applications requiring high spatial/temporal resolutions such as from finance, which is also one of the fastest-growing sectors where information is data type agnostic, according to Earth Observation Data and Services Market and Earth Observation Satellite Systems. With spatial resolution in the 5-10m range, this could be considered suitable for most resource extraction and management applications such as agriculture, forestry, oil and gas, and mining. With sharper resolutions (spatial and spectral), security applications could also drive demand. Defense being the most promising market, there is a clear sensitivity in using hyperspectral data to detect true and camouflaged objects on a very responsive change detection basis.
