GMV delves into high-performance onboard processing for vision-based autonomous space navigation
On September 27, the final review meeting was held for the GMVision project (On-board image processing architecture and co-processor upgrade for vision-based navigation), a European Space Agency project, which aims to redesign, develop, and validate an engineering model (EM) of a high-performance on-board co-processing unit that redundantly processes images from two cameras for different autonomous vision-based navigation scenarios for aircraft and space exploration satellites.
The space aviation electronics system consists of two navigation cameras and a highly versatile redundant image processing co-processor (IPB) that could be used for rendezvous and search operations for satellites, as well as for descents and landings on asteroids, moons, and other small celestial bodies.
The solution resulting from this activity was developed in its entirety by GMV, from the concept to the software and then on to the development of the camera hardware and processing electronics. On the software side, the proposed solution is designed for complex vision-based algorithms, where core parts of the software solution are hardware accelerated to run in parallelizable solutions.
The system was tested following a complete validation chain that included proof-of-concept testing, a black room testbed, and system testing with real mock-ups.