Metaspectral to Partner with Armada on Edge AI

Vancouver-based Metaspectral has announced a new partnership with edge-computing company Armada. The companies will be working together on incorporating Metaspectral’s hyperspectral imaging technology to Armada’s edge computing marketplace, with a focus on Armada’s “Edge AI.” SpaceQ reached out to Metaspectral for more details. CEO and co-founder Francis Doumet provided answers to our questions. Metaspectral and Hyperspectral Imaging As mentioned in previous SpaceQ coverage, Metaspectral’s focus is on hyperspectral imaging. Unlike traditional optical images that have three spectral bands for each pixel (corresponding to red, green, and blue), each pixel in a hyperspectral image can have complete spectrums of hundreds of different wavelengths per pixel. Thus, each hyperspectral image is actually more like a data cube: the x dimension, y dimension, and a third “dimension” consisting of each pixel’s information on those hundreds of different wavelengths. Those images can then be analyzed, particularly by machine learning AI, to determine useful insights. Satellite-based hyperspectral imaging is extremely useful for everything from search and rescue to agriculture and meteorology. In particular, it can be useful for tracking and predicting wildfires, which is why Metaspectral created a team for the XPRIZE Wildfire competition to develop tools and technology aimed at identifying wildfires. The company also has a planned deployment of a payload to the International Space Station, to analyze hyperspectral data in real time. Nevertheless, hyperspectral imagery is enormous, with a single hyperspectral image ranging up to four gigabytes in size. This can make it a tremendous challenge to transmit, considering the comparatively-limited bandwidth that is usually available over satellite-based connections in remote locations. One of the key capabilities that Metaspectral developed back in 2022 was a way to compress that data using custom FPGA hardware; losslessly compressed at up to 40% of the original size, and near-losslessly at up to 10% of the original size. Even so, edge computing—where the AI analysis of the bulky imagery is done locally—is a logical pairing, as it would ensure that only the analyses would need to be transmitted. Armada’s Edge Computing Galleons Armada might help provide a solution. Armada offers remote Starlink connection solutions, which could help with transmission issues, but their more relevant technology in this case is their “Galleon” self-contained datacenters. Armada describes them as “rapidly deployable, containerized modular data centers;” large edge computing platforms sized like (and reminiscent of) cargo shipping containers. They’re built to be carried by trucks and trains much like a cargo container, and to be deployed in remote locations: like oil rigs, mining sites, agricultural operations, and possibly even conflict zones. The Galleon datacenters are then monitored and controlled via remote Starlink-based connections. And, when deployed, one of the main applications is AI, whether generative AI or (as in Metaspectral’s case) analytical machine learning algorithms. Armada’s term for it is “EdgeAI”, which is focused on “efficient deployment of compute-intensive, latency-sensitive applications at the edge,” including “live video interpretation, inspection, monitoring, data stream aggregation, and more.” These applications are hosted on an App Marketplace, similar to the App Stores found on a smartphone. According to the announcement, that’s where Metaspectral comes in. “Metaspectral’s advanced computer vision capabilities will be integrated into Armada’s Edge AI Marketplace,” the company said, adding that its “proprietary data compression algorithms also enable real-time analysis and transmission from satellite and terrestrial sources without compromising image quality.” Doumet added in his comments to SpaceQ that “Armada customers will have the ability to purchase Metaspectral's hyperspectral analysis software directly from Armada's marketplace, and use it to begin analyzing hyperspectral data with AI in real-time.” “Synergistic companies” In his comments to SpaceQ, Doumet emphasized how “Armada and Metaspectral are synergistic companies.” “Armada provides ruggedized compute infrastructure designed to operate in remote areas,” he said, while “Metaspectral provides sensing hardware and AI to run on the hardware.” Each company benefits from the others’ capabilities: “Metaspectral gets access to ruggedized hardware, complete with a Starlink connection for connectivity in remote areas,” while “Armada gets to enter the industrial computer vision market with Metaspectral's technology.” He also said that the Starlink connections will be key in enabling Metaspectral to “push analysis results to the cloud for consumption by the customer,” showing that LEO constellation access is still an important factor even in terrestrial edge AI applications. Potential customers, according to Doumet, include “mining, oil & gas, and defence.” Mining customers can remotely “sample the grade of ore coming out of a mine,” defence customers will gain enhanced “situational awareness,” and oil & gas customers will be able to monitor methane leaks. Those and other customers will have access relatively quickly: Doumet said that “Metaspectral's software has already been ported to run on Armada hardware…[and] will be available on the Marketplace in a matter of weeks.” In the announcement, Metaspectral CTO (and co-founder) Mugel Tissera also discussed potential applications, saying that hyperspectral imagery “can be used in the oil and gas sector to detect pipeline leaks, monitor vegetation health, and identify hydrocarbon reservoirs,” while mining companies can use it “for mineral identification, prospecting support, and monitoring environmental impacts such as soil erosion, vegetation health, and water quality.” He also pointed to “extensive” defence applications, “including surveillance, detection, and identification of hidden objects, terrain assessment, and more.” In turn, Uday Tennety, VP of Product Management at Armada, said that Armada sees “a tremendous opportunity to help organizations unlock the power of real-time insights and advanced analytics to drive business growth and innovation” coming from the partnership. He said that Armada looks forward to “ensuring our joint customers across critical industries are able to glean actionable insights from their data, optimize processes to prioritize agility and precision, and streamline operations.”