Deep Learning in Remote Sensing
Co-organised by IEEE Geoscience and Remote Sensing Society and ISPRS
As in many other fields deep learning has made major inroads in the interpretation of remote sensing imagery during the last few years. For image interpretation, the recent revolution in performance started in computer vision with the presentation of AlexNet in 2012, reducing the errors in the ImageNet Large-Scale Visual Recognition Challenge to half of its previous values. Since then, deep learning approaches have outperformed classical methods in just about each and every field in computer vison, and success has also been reported in photogrammetry and remote sensing.
In this thematic session, co-organised by the two major international scientific societies in the field, IEEE GRSS and ISPRS, the current state-of-the-art and future directions of this vibrant field are presented.
Paolo Gamba (University of Pavia) and Christian Heipke (Leibniz University Hannover)
Suggested presenters (sequence subject to change) and working titles:
Xiaoxiang Zhu (TU Munich) – Deep learning for mapping global urban morphology from space
Begüm Demir (TU Berlin) – Deep Hashing for Scalable Remote Sensing Image Retrieval in Large Archives
Ribana Roscher (Uni Bonn) – Scientific Domain Knowledge for Explainable Machine Learning
Devis Tuia (Wageningen University) – Geospatial machines interacting with people