Since its very beginnings, the extraction of information from remote sensing data has witnessed three clearly distinct stages: visual interpretation, parametric methods and, more recently, non-parametric approaches. Within the domain of the latter one, deep learning evolved as a broader group of machine learning resources and emerges as a new paradigm targeted to data-intensive science. Besides remote sensing, it has encountered applications in the most diverse fields, which range from computer vision, speech and audio recognition, medical image analysis to social network filtering, bioinformatics and game programs. Deep learning meets the big data processing demands in remote sensing, is a fast growing trend in image analysis, still far from being exhausted, and fits well into the current needs of smart and near real time applications.
The ever-increasing number of Earth observation satellites currently orbiting our planet daily delivers huge amounts of data, with an order of magnitude of over a hundred petabytes. In parallel to this, the growing use of low altitude platforms, with extremely high spatial resolution, generate as well massive data for digital processing. We are then faced with the requirements of robust and computational cost-effective algorithms for handling high-dimensional datasets in an efficient manner. Deep learning arises as a promising alternative for tackling these challenges and has demonstrated to be able to deal with bulky and noisy spectral and spatial information. It has attracted the attention of the practitioners and scientific communities and is a steadily growing field of research in remote sensing. This Research Topic is committed to report the latest advances in deep learning approaches applied to remote sensing and welcomes theoretical and theory application papers. We truly expect that the contributions to this Research Topic may be a benchmark for those engaged with the continuous advancement of remotely sensed images processing and interpretation. Articles may included the following topics, but are not limited to:
• Deep Learning for Image Processing (denoising, segmentation, classification)
• Deep Learning for Image Interpretation (semantic labeling, object detection, image indexing)
• Deep Learning for Data Fusion (multisensor analyses, pansharpening)
• Innovative Approaches for Data Augmentation
• Deep Learning and GEOBIA
• Deep Learning for Time Series Filtering and Dimensionality Reduction
• Extreme Learning, Transfer Learning, Cross-sensor Learning
• Deep Learning for 3D Reconstruction
• New Architectures of Deep Learning
Since its very beginnings, the extraction of information from remote sensing data has witnessed three clearly distinct stages: visual interpretation, parametric methods and, more recently, non-parametric approaches. Within the domain of the latter one, deep learning evolved as a broader group of machine learning resources and emerges as a new paradigm targeted to data-intensive science. Besides remote sensing, it has encountered applications in the most diverse fields, which range from computer vision, speech and audio recognition, medical image analysis to social network filtering, bioinformatics and game programs. Deep learning meets the big data processing demands in remote sensing, is a fast growing trend in image analysis, still far from being exhausted, and fits well into the current needs of smart and near real time applications.
The ever-increasing number of Earth observation satellites currently orbiting our planet daily delivers huge amounts of data, with an order of magnitude of over a hundred petabytes. In parallel to this, the growing use of low altitude platforms, with extremely high spatial resolution, generate as well massive data for digital processing. We are then faced with the requirements of robust and computational cost-effective algorithms for handling high-dimensional datasets in an efficient manner. Deep learning arises as a promising alternative for tackling these challenges and has demonstrated to be able to deal with bulky and noisy spectral and spatial information. It has attracted the attention of the practitioners and scientific communities and is a steadily growing field of research in remote sensing. This Research Topic is committed to report the latest advances in deep learning approaches applied to remote sensing and welcomes theoretical and theory application papers. We truly expect that the contributions to this Research Topic may be a benchmark for those engaged with the continuous advancement of remotely sensed images processing and interpretation. Articles may included the following topics, but are not limited to:
• Deep Learning for Image Processing (denoising, segmentation, classification)
• Deep Learning for Image Interpretation (semantic labeling, object detection, image indexing)
• Deep Learning for Data Fusion (multisensor analyses, pansharpening)
• Innovative Approaches for Data Augmentation
• Deep Learning and GEOBIA
• Deep Learning for Time Series Filtering and Dimensionality Reduction
• Extreme Learning, Transfer Learning, Cross-sensor Learning
• Deep Learning for 3D Reconstruction
• New Architectures of Deep Learning
