Dealing with multiple sources


Deep nets can be used efficiently to deal with multiple heterogeneous sources, for instance, in implementing a number of branches tailored for each data source, that eventually merge together at some point to fuse the information. For example, it is easy to build a deep neural network that can deal simultaneously with time series and one very-high-resolution image: while one branch composed of a recurrent neural network processes the time series (For instance Sentinel-1 or Sentinel-2 images), another branch composed of a CNN processes the high-resolution image (for instance Spot 6/7 images). Such approaches are available in the literature (for instance in [13]).

This section will present how OTBTF can be ...

Get Deep Learning for Remote Sensing Images with Open Source Software now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.