Hearing about the launch of the French military CSO 2 optical reconnaissance satellite introduced in the first post of this blog series, took me back to the early 2000’s when our technology was included into the NASA DAAC (Distributed Active Archive Center). Back then, the challenge was to provide technology with the ability to ensure wide distribution of vast amounts of remote sensing data for analysis and exploitation using open standards.
Since then, and with the cost of launching satellites getting lower and lower, the sky has been invaded by all types of earth observation satellites, expanding the initial challenge and contributing to what we call the remote sensing data deluge. In recent years, low-orbit satellite constellations like the French CSO have been added to the mix making even more data available and providing revisit times like never before. It is now possible to have a daily update of satellite data on every point of the planet.
This huge amount of up-to-date data has tremendous value for imagery intelligence (IMINT), including the ability to operate at a distance, removed from physical boundaries, with a technology (satellites) that is out of reach of enemy air defenses. In most cases, these capabilities are critical for a wide range of military decision-making.
However, this explosion of data availability combined with a lack of resources within defense departments trained to maximize the value of this data, requires a new approach based on more automation and the use of artificial intelligence and machine learning.
The Hexagon technology used back in the days of the NASA DAAC matured to become ERDAS APOLLO, which is able to serve vast amounts of data, including satellite data in a secured way directly to the applications for decision making. One of the key features of ERDAS APOLLO is reducing human intervention by using automation to catalog, secure, and publish data through standard interfaces like OGC, ISO, REST API, and others. ERDAS APOLLO also has a geoprocessing service based on ERDAS IMAGINE’s Spatial Modeler that includes processing for object recognition or change detection, which is enhanced with artificial intelligence using machine learning or deep learning techniques. The geoprocessing service allows for further data exploitation through dedicated web client applications like M.App X or Catalog Explorer.
From its 480 km orbit, the CSO 2 satellite will provide EHR (Extremely High Resolution) imagery (around 20 cm resolution) as well as 3D imagery in the visible and infrared bands, allowing data capture day and night for maximum exploitation. Thanks to LuciadFusion even 3D data can be used in M.App X or Catalog Explorer to improve intelligence extraction by providing a more realistic and recognizable context to image analysts.
The current Hexagon technology stack provides very impressive tools for intelligence exploitation of CSO 2 – and other remote sensing – data. But I expect to see even more advanced applications in the future as we continue to integrate more automation and artificial intelligence into our technologies. So, stay tuned!