This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. For attribution, the original author(s), title, publication source (PeerJ Preprints) and either DOI or URL of the article must be cited.
Cite this article
Arellano-Verdejo J, Lazcano-Hernandez H, Cabanillas-Terán N.2018. ERISNet: Deep learning network for Sargassum detection along the coastline of the Mexican Caribbean. PeerJ Preprints6:e27445v1https://doi.org/10.7287/peerj.preprints.27445v1
Recently, Caribbean coasts have experienced atypical massive arrivals of pelagic Sargassum with negative consequences both ecologically and economically. Based on deep learning techniques, this study proposes a novel algorithm for floating and accumulated pelagic Sargassum detection along the coastline of Quintana Roo, Mexico. Using convolutional and recurrent neural networks architectures, a deep learning network (named ERISNet) was designed specifically to detect this macroalgae along the coastline through remote sensing support. A new dataset which includes pixels values with and without Sargassum was built to training and testing ERISNet. Aqua-MODIS imagery was used to build the dataset. After the learning process, the designed algorithm achieves a 90 % of probability in its classification skills. ERISNet provides a baseline for automated systems to accurately and efficiently monitor algal blooms arrivals.
This is a submission to PeerJ for review.
Training dataset for proposed algorithm
The dataset included 14 different attributes and 4515 instances, of which 2306 corresponded to the presence of Sargassum and 2209 without. Dataset was built with the pixel data of each MODIS-band for different dates.