Develop a raw image compressor achieving metrological quality, high speed and high compression ratio. Compressed images must remain are suitable for any post-processing task, including advanced AI algorithms. The target of the project is to implement the compressor for space applications, as software, VHDL code and on an Intel Movidius VPU.

Artificial intelligence will allow vast quantities of earth observation data to be analyzed. Image compression can both provide a performance boost and make systems more economical. However, most high-performance compression methods have been tuned to the human eye, rather than developed for machine vision, generating artefacts that although invisible limit the performance of advanced algorithms. Compression targeting AI and high radiometric accuracy applications is needed.

Results confirmed the raw quality preservation and a compression ratio in between 5:1 and 10:1, with high throughput and low complexity / energy use. This could reduce costs for data transfer and data storage on ESA Earth Observation programme by a factor of 3.

Dotphoton is applying for new funding through ESA GSTP E2 Make in order to achieve TRL5 for on-board compression and ground-based decompression, starting in Q1 2021. The main markets for this new development are space, biomedical and automotive.