Today’s satellites are inherently inflexible and purpose-built for a single mission, relying heavily on the ground segment to analyze data. The space market is looking for on-board flexibility / on-board reconfiguration targeting the digitalization of payload for earth observation and space exploration missions. Furthermore, there is an increasing interest for the availability of AI-oriented hardware devices allowing tasks such as deep learning inference and pre-processing of sensor data in-orbit.

OHB is constantly exploring disruptive approaches to propose its customers always better space-based solutions. From its existing heritage in the domain, the company is exploring the possibility to transfer part of the processing of satellite data from the earth (ground segment) to space (space segment), through the development of so-called On-board Payload Data Processing (OPDP).

Some example of applications:

  • Data fusion on board
  • Automatic optical sensor calibration
  • End-to-End secured data transfer
  • Filtering onboard the data of interest increasing the flexibility in usage of bandwidth, data storage and coverage.

The technology can be an enabler for new applications, for example:

  • In-orbit distributed calculation
  • fleet cooperation
  • automation of mission control and operation

We are currently examining the possible combination of software framework and hardware from the use of AI accelerator to high-performance mixed-criticality avionics electronics. Our work includes design mitigation techniques to include this development in existing OHB satellite platform.

Based on the existing group heritage and competences concerning onboard data processing, OHB Hellas is strongly committed to providing a turn-key system with a combination of AI-oriented hardware and open software to bring intelligence on board while providing a high level of flexibility. This solution would enhance the group portfolio and further improve its competitiveness for future earth observation and space exploration missions.

Civil Security from Space

Through high-performance on-board data processing, the response time to critical events can be reduced, as data is available to decision makers faster. Applications of civil security are critical for Greece, given the catastrophic phenomena that occur frequently, such as wildfires and given the country's sensitive geographic positioning as well. High performance on board data processing, coupled with technologies such as virtualization and flexible ground stations can provide innovative platforms that are part of a distributed network. This network can benefit remote sensing applications from space which can be implemented to answer Greece's needs and create a sovereign product at the same time, as the heritage already exists in the country.

Onboard Multi-Frame Super Resolution Image study

Super-resolution refers to the process of improving the spatial resolution, that is, the level of detail, of an image or an acquisition system. Our innovative idea is to use the OPS-SAT powerful on-board computer to deploy a machine learning-based super-resolution algorithm directly on-board the satellite. It has the potential of opening new perspectives of EO applications mainly in time-critical domains like safety and security. In the frame of this experiment, we successfully developed and deployed an AI-based multi-frame super resolution algorithm on board OPS-SAT achieving a notable improvement in image quality and proving the merits of the proposed approach. 

Cognitive Cloud Dual Camera study

The main objective of this study undertook by OHB Hellas is the evaluation of the merits of a dual camera setup for the acquisition of both large swath, low- resolution images and narrow-swath, high-resolution images of identified events of interest, on-board, through AI methods. The feasibility of the approach on a mission-level concept of a single-satellite, dual camera setup as well as a two-satellite (leading-trailing), single camera setup was investigated. The AI approach was demonstrated on commercially available representative high-performance on-board computers (Xilinx Kria SoM / Google Edge TPU SoM) through a state-of-the-art fire detection use case.