Engineers from the Gliwice are working on new image processing methods that will increase image resolution for satellites and revealing previously unseen details.
The project from KP Labs called 'Deep Sent', which has been selected for funding by the European Space Agency (ESA), will enable wider use of images from earlier satellites, for example for land mapping, spatial planning, decision support in the insurance industry or quick response to crisis situations in order to protect the environment. Resolution enhancement may also find applications in radiology, for magnifying computed tomography images.
Engineers will use super-resolution reconstruction to increase the resolution of images and obtain previously unseen details instead of a pixel grid. To achieve this, algorithms will be used to select data for machine learning. Engineers will first use existing networks for super-resolution reconstruction of Sentinel-2 images in such a way that each band of the multispectral image will be processed independently, and then attempt to exploit the correlation between each band.