AI4Space: Transfer Learning for On-Orbit Ship Segmentation

The Ubotica team are looking forward to participating in this year’s virtual AI4Space workshop which takes place on 23rd October.

The workshop will highlight the space capabilities that draw from and/or overlap significantly with vision and learning research, outline the unique difficulties presented by space applications to vision and learning, and discuss recent advances towards overcoming those obstacles.

The general emphasis of AI4Space is vision and learning algorithms in off-Earth environments, including in the orbital region, surface and underground environments on other planetary bodies, interplanetary space and solar system, and distant galaxies.  Target application areas include autonomous spacecraft, space robotics, space traffic management, astronomy, astrobiology and cosmology. Emphasis is also placed on novel sensors and processing hardware for vision and learning in space, mitigating the challenges of the space environment towards vision and learning and solving practical difficulties in vision and learning for space.

Vincenzo Fanizza, Ubotica’s Space Systems Engineer will be giving a spotlight presentation about ‘Transfer Learning for On-Orbit Ship Segmentation’.  The presentation will focus on the new artificial intelligence on Ubotica’s satellites which is being used to identify and track ships on the oceans.  Vincenzo will explain how Ubotica’s AI technology will deliver faster and more accurate images of ships for the end users.

“We train AI with images processed to match the ones at the stage of on-orbit inference. Ubotica will be able to capture better quality information in orbit related to any ship anywhere in the world.” Vincenzo notes.