Research themes
GCER lab offers an interdisciplinary space for agricultural and biosystems engineers, computer scientists, software developers, and geographers to engage in Earth science, deep learning, and new remote sensing technologies. Our research themes are centered on land and water resources monitoring using AI-driven techniques and satellite data.
Deep learning for terrestrial monitoring
Deep learning techniques support innovations in Land Resources Essential Information Systems (LREIS) using high spatial resolution satellite imagery, and the next generation of land cover and land change products will provide essential information for assessing and managing land and natural resources in higher spatial resolution and near-real time.
Related publications:
-
Change detection for burned area mapping: Martins et al. (2022) https://doi.org/10.1016/j.rse.2022.113203
-
Object-based CNN 1-m Land Cover: Martins et al. (2020) https://doi.org/10.1016/j.isprsjprs.2020.08.004
-
Neural network 1.2-m wetland mapping: Martins et al. (2020) https://doi.org/10.1016/j.jag.2020.102215
Remote sensing of Water resources
Satellite imaging systems have been widely used to support water quality and quantity management as it allows water quality analysis using empirical and semi-analytical relationship between optically active constituents (e.g., sediments, chlorophyll-a, dissolved organic matter) and spectral water reflectance. This is particularly relevant for climate-sensitive regions that lack water quality stations and have many extreme climate events that can disturb the aquatic environment.
Related publications:
-
Remote sensing of large reservoir in the drought years: Martins et al. (2019). https://doi.org/10.1016/j.rsase.2018.11.006
-
Atmospheric correction of satellite images for water: Martins et al. (2017) https://doi.org/10.3390/rs9040322
-
Adjacency correction for inland waters: Paulino et al. (2022) https://doi.org/10.3390/rs14081829
-
Diffuse attenuation coefficient in optically complex waters: Maciel et al. (2020). https://doi.org/10.1016/j.isprsjprs.2020.10.009
Gallery: work & people
