Zoeken
Zoeken kan via de modus 'eenvoudig zoeken' (één veld) of uitgebreid via 'geavanceerd zoeken' (meerdere velden). Zo kan je bv. zoeken op een combinatie van een auteursnaam (auteur), een jaartal (jaar) en een documenttype.
Boekenmand
Nuttige resultaten kan je aanvinken en toevoegen aan een mandje. De inhoud hiervan kan je exporteren of afdrukken (naar bv. PDF).
RSS
Op de hoogte blijven van nieuw toegevoegde publicaties binnen uw interessegebied? Dit kan door een RSS-feed (?) te maken van jouw zoekopdracht.
nieuwe zoekopdracht
DINCAE 2.0: multivariate convolutional neural network with error estimates to reconstruct sea surface temperature satellite and altimetry observations
Barth, A.; Alvera-Azcárate, A.; Troupin, C.; Beckers, J.-M. (2022). DINCAE 2.0: multivariate convolutional neural network with error estimates to reconstruct sea surface temperature satellite and altimetry observations. Geosci. Model Dev. 15(5): 2183-2196. https://dx.doi.org/10.5194/gmd-15-2183-2022
In: Geoscientific Model Development. Copernicus Publications: Göttingen. ISSN 1991-959X; e-ISSN 1991-9603, meer
|   |
| Auteurs | | Top |
- Barth, A.
- Alvera-Azcárate, A.
- Troupin, C.
- Beckers, J.-M.
|
|
|
| Abstract |
DINCAE (Data INterpolating Convolutional Auto-Encoder) is a neural network used to reconstruct missing data (e.g., obscured by clouds or gaps between tracks) in satellite data. Contrary to standard image reconstruction (in-painting) with neural networks, this application requires a method to handle missing data (or data with variable accuracy) already in the training phase. Instead of using a standard L2 (or L1) cost function, the neural network (U-Net type of network) is optimized by minimizing the negative log likelihood assuming a Gaussian distribution (characterized by a mean and a variance). As a consequence, the neural network also provides an expected error variance of the reconstructed field (per pixel and per time instance). In this updated version DINCAE 2.0, the code was rewritten in Julia and a new type of skip connection has been implemented which showed superior performance with respect to the previous version. The method has also been extended to handle multivariate data (an example will be shown with sea surface temperature, chlorophyll concentration and wind fields). The improvement of this network is demonstrated for the Adriatic Sea. Convolutional networks work usually with gridded data as input. This is however a limitation for some data types used in oceanography and in Earth sciences in general, where observations are often irregularly sampled. The first layer of the neural network and the cost function have been modified so that unstructured data can also be used as inputs to obtain gridded fields as output. To demonstrate this, the neural network is applied to along-track altimetry data in the Mediterranean Sea. Results from a 20-year reconstruction are presented and validated. Hyperparameters are determined using Bayesian optimization and minimizing the error relative to a development dataset. |
IMIS is ontwikkeld en wordt gehost door het VLIZ.
