Intel chip powers PhiSat-1, the first satellite with an onboard AI processor, opening the door to the age of smart satellite imaging

November 20, 2020 | Case Study

The performance of cameras is increasing at a rapid pace; every new-generation camera captures approximately 100 times more data than the previous one. However, the performance of satellite communication equipment for data transmission is increasing at a lower rate, by a factor of up to five per generation. This causes a bottleneck in the capabilities of satellite imaging as images are downlinked and processed on the Earth station. In addition, two-thirds of our planet's surface is covered by clouds at any given time, so collecting and transmitting images occluded by clouds is often a total waste. To solve this problem, startups including Orbital Sidekick are attempting to include edge computing functionalities on satellites to process the data on board. However, power requirements and the accuracy of edge computing remain barriers.