Moderate resolution imaging spectroradiometer products classification using deep learning
During the last years, the algorithms based on Artificial Intelligence have increased their popularity thanks to their application in multiple areas of knowledge. Nowadays with the increase of storage capacities and computing power, as well as the incorporation of new technologies for massively parallel processing (GPUs and TPUs) and Cloud Computing, it is increasingly common to incorporate this kind of algorithms and technology in tasks with a deep social and technological impact. In the present work a new Convolutional Neural Network specialized in the automatic classification of Moderate Resolution Imaging Spectroradiometer satellite products is proposed. The proposed architecture has shown a high-generalization by classifying more than 250,000 images with 99.99% accuracy. The methodology designed also can beextended, with other types of images, to make detection of Sargassum, oil spills, red tide, etc.
| Main Author: | |
|---|---|
| Format: | Texto biblioteca |
| Language: | eng |
| Subjects: | Sensores remotos, Aprendizaje profundo, Aprendizaje automático (Inteligencia artificial), Algoritmos, |
| Online Access: | http://dx.doi.org/10.1007/978-3-030-33229-7_6 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|