The goal of the DeepFlow project was to train a deep neural network with large pools of data and find similar patterns in thousands of new input images.
DeepFlow helps to collect information from satellite pictures. It will allow to train a deep neural network with large pools of data and find similar patterns in thousands of new input images. If the images require pre-processing and further transformations, these can be performed by DeepFlow image correction mechanisms. Additionally, thanks to the integration of the DeepFlow framework with the Copernicus programme through the DIAS platform, it is possible to access free-to-use images captured by Sentinel and Landsat satellites. All these features combined make it possible to acquire meaningful information in a much shorter time. With DeepFlow you will be able to redefine the way you gather data for your organization and get access to new, so far unavailable means of data analysis.