Aquaculture and saliculture are relevant economic activities in the Brazilian Coastal Zone (BCZ). However, automatic discrimination of such activities from other water-related coverages / uses is not an easy task. In this sense, convolutional neural networks (CNN) have the advantage of predicting the class label of a given pixel, providing as input a local region (patches or named chips) around that pixel. Both the convolutional nature and the semantic segmentation capability provide the U-Net classifier with the ability to access the "context domain" instead of just isolated pixel values. Supported in the context domain, we present the results of the analyzes.
Here you can find and learn more about the scientific articles prepared by the MapBiomas team, they are related to issues involving the mapping of land use in Brazil.
In this study, we developed an alternative approach for mapping burned areas in the Cerrado biome in Brazil, using Landsat imagery and Deep Learning algorithm, implemented on the Google Earth Engine and on Google Cloud Storage platform.
Understanding the dynamics of native forest loss and gain is critical for biodiversity conservation and ecosystem services, especially in regions experiencing intense forest transformations. We quantified native forest cover dynamics on an annual basis from 1990 to 2017 in Brazil’s Atlantic Forest.
In this study, we described a novel approach and the results achieved by a multi-disciplinary network called MapBiomas to reconstruct annual land use and land cover information between 1985 and 2017 for Brazil, based on random forest applied to Landsat archive using Google Earth Engine.
In this study, we utilized the entire set of Landsat 8 images available for Brazil in 2015, from which dozens of seasonal metrics were derived, to produce, through objective criteria and automated classification strategies, a new pasture map for the country.
In this work, we construct a model to evaluate the possible consequences of policy actions on land cover dynamics in the near-future at a high-resolution scale.
The Brazilian Cerrado represents the largest savanna in South America, and the most threatened biome in Brazil owing to agricultural expansion. To assess the native Cerrado vegetation (NV) areas most susceptible to natural and anthropogenic change over time, we classified 33 years (1985–2017) of Landsat imagery available in the Google Earth Engine (GEE) platform.
In this paper, we propose a method to automatically detect and map center pivot irrigation systems using U-Net, an image segmentation convolutional neural network architecture, applied to a constellation of PlanetScope images from the Cerrado biome of Brazil. Our objective is to provide a fast and accurate alternative to map center pivot irrigation systems with very high spatial and temporal resolution imagery.
This study evaluated, based on thousands of PlanetScope images obtained over a 12-month period, the performance of three machine learning approaches (random forest, long short-term memory-LSTM, and U-Net). We applied these approaches to mapped pasturelands in a Central Brazil region.
In this study, we present a long-term spatiotemporal analysis of surface water annual change and address potential connections with deforestation, infrastructure expansion and climate change in this region.
The differences found in the products point to distinct and complementary uses, with the Atlas base most suitable for identifying deforestation and analyzing biodiversity conservation and MapBiomas for analysis of fragmentation, restoration and water protection. The option for the product that will be used varies according to the objective of the proposed work.
This work aimed to report the experiences of the application of a didactic workshop that dealt with the potential of the products and tools of the Mapbiomas Project in a class of the Graduate Program in Modeling in Earth and Environmental Sciences, from the State University of Feira de Santana.
This work has mapped, annually and in an unprecedented way, the totality of the Brazilian pastures, from 1985 to 2017. With an overall accuracy of about 90%, the 33 maps produced indicated the pasture area varying from ~118 Mha ±3.41% (1985) to ~178 Mha ±2.53% (2017), with this expansion occurring mostly in the northern region of the country and to a lesser extent in the midwest.
This study mapped, through objective criteria and automatic classification methods (Random Forest) applied to MODIS (Moderate Resolution Imaging Spectroradiometer) images, the totality of the Brazilian pastures between 2000 and 2016.
In this study, methods, originally developed to assess life course trajectories, are explored in order to evaluate land change through the analysis of sequences of land use/cover. Annual land cover maps which describe land use/land cover change for the 1985–2017 period for a large region in Northeast Brazil were analyzed.
This manuscript presents a Google Earth Engine (GEE)-managed pipeline to compute the annual status of Brazilian mangroves from 1985 to 2018, along with a new spectral index, the Modular Mangrove Recognition Index (MMRI), which has been specifically designed to better discriminate mangrove forests from the surrounding vegetation. If compared separately, the periods from 1985 to 1998 and 1999 to 2018 show distinct mangrove area trends.