S5P_SISR_Toolbox v1.1.0

Version 1.1.0 represents the time-saving use of super-resolution methods on Sentinel-5P Level-1b data.

The algorithms are tested on three images consisting of 3445 channels, once they have been pre-processed.

The algorithms available are: interpolation (cubic); non-blind deconvolution solved with CGA; DL-based methods (some SOTA, our original S5Net fine-tuned independently for each channel, and our original S5Net fine-tuned with a newly developed multidirectional cascade framework).

S5P_SISR_Toolbox v1.0.0

Version 1.0.0 represents the first use of super-resolution methods on Sentinel-5P Level-1b data.

The algorithms are tested on eight monochromatic images taken from two distinct orbits, corresponding to one channel for each detector. The images were simply cropped along- and across- track, with no additional pre-processing.

The algorithms available are: interpolation (cubic); non-blind deconvolution solved with CGA; DL-based methods (some SOTA and our original S5Net fine-tuned independently for each channel).