Deep Learning for Environmental Applications of Remote Sensing Data
A Special Issue from Canadian Journal of Remote Sensing
Read this page in FrenchAbout this Special Issue
Inspired by the great potential of human brains for object recognition, Deep Learning (DL) has drawn attention within the remote sensing community over the past few years. Supervised convolutional neural network (CNN), recurrent neural network (RNN), unsupervised Auto-Encoders (AE), deep belief network (DBN), and generative adversarial network (GANs) are the state-of-the-art DL methods applied for remote sensing imagery. Given a large enough training set, these methods are advantageous compared to conventional shallow structured machine learning tools, such as neural networks (NN), support vector machines (SVM), and ensemble algorithms, e.g., random forest (RF), which have been successfully used in the analysis of remotely sensed data for several years.
The popularity of DL methods is attributed to both their deep multilayer structure, allowing the extraction of robust, invariant, and high-level features of data, and to their end-to-end training scheme. In other words, these methods have the capability to learn a series of abstract hierarchical features from raw input data, and to provide a final, task-specific output, thus removing heuristic feature engineering design.
The Canadian Journal of Remote Sensing aims at publishing a special issue on “Deep Learning for Environmental Applications of Remote Sensing Data.” The main objective of this special issue is to promote the recent thematic research and development applications of deep learning approaches for a variety of remote sensing problems. Papers of applicative nature providing new deep learning-oriented public datasets for the remote sensing community are welcome.
Submissions are encouraged to cover a broad range of topics such as transfer learning, design of new deep learning architectures, efficient training of deep learning architectures, preparing large-scale deep learning-oriented public dataset, deep reinforcement learning, deep learning model in Geo Big Data, which may include, without being limited to, the following applications: image fusion, segmentation and classification; pan-sharpening, denoising and super-resolution; disaster responses (e.g., oil spill, inundation); land use/land cover change detection; target detection (e.g., ship and iceberg); crop yield prediction, and environmental monitoring (e.g., wetland and forest).
Submissions Information
All manuscripts will be subjected to peer-review, can be submitted in either French or English and are subject to standard publication charge (a total of $400 USD for a research or review paper, $300 USD for a research note).
Prospective authors should follow the regular guidelines of the CJRS, see Instructions for Authors. Electronic submissions following CJRS guidelines should be submitted via ScholarOne Manuscripts by June 30th 2020.
Targeted publication date will be early of 2021.
Click on the options below to read the Aims & Scope, Instructions for Authors and to Submit your article now.
Contact
CJRS-JCT Editor-in-Chief: Monique Bernier, INRS – Québec City, QC, Canada
Guest Editors
Masoud Mahdianpari Memorial U. / C-Core St. John's, NL, Canada
Saeid Homayouni INRS Quebec, Qc, Canada
Samuel Foucher CRIM Montreal, Qc, Canada