A dynamic and generic cloud computing model for glaciological image processing

• We use in situ pictures to complement satellite imagery.
• We process the pictures to map the temporal evolution of the glacier snow cover.
• We build a cloud computing model to process the in situ pictures.

As satellite imaging is quite expensive, and because of poor weather conditions including common heavy cloud cover at polar latitudes, daily satellite imaging is not always accessible or suitable to observe fast temporal evolutions. We complement satellite imagery with a set of ground based autonomous automated digital cameras which take three pictures a day. With these pictures we build a mosaic with their projection and apply a classification to define the temporal evolution of the snow cover. As the pictures are subject to heavy disturbance, some processing is needed to build the mosaic. Once the processes are defined, we present our model. This model is built upon a cloud computing environment using Web services workflow. Then we present how the processes are dynamically organized using a scheduler. This scheduler chooses the order and the processes to apply to every picture to build the mosaic. Once we obtain a mosaic we can study the variation of the snow cover.