A current finding of TB13 EOC is that job mapping requirements for EO data will mainly be conducted as batch processing of spatially unrelated images. Also, another finding indicates that computing ressources available (VM flavors) will be of adequate size to process the individual input images. This severly restrains the use case scenarios.
A more general use case would take in account full imagery swaths with overlapping information between individual images. An advanced job mapper would support a better estimation of required ressources for the job, in terms of number of tiles to process the swath, optimal size of tiles according to VM specs, use of supporting optical imagery as prior information, as well as reducing the outputs into a single coherent one.
A current finding of TB13 EOC is that job mapping requirements for EO data will mainly be conducted as batch processing of spatially unrelated images. Also, another finding indicates that computing ressources available (VM flavors) will be of adequate size to process the individual input images. This severly restrains the use case scenarios.
A more general use case would take in account full imagery swaths with overlapping information between individual images. An advanced job mapper would support a better estimation of required ressources for the job, in terms of number of tiles to process the swath, optimal size of tiles according to VM specs, use of supporting optical imagery as prior information, as well as reducing the outputs into a single coherent one.
Proposed sponsor: NRCan