This repository contains samples used in the Restore+ Project to produce the Land Use and Land Cover (LULC) maps for Brazil. These samples were built by combining and processing base samples from Brazilian national agencies and scientific collaborations. All of them cover multiple Brazilian regions and distinct time periods.
The samples are available in ready-to-use formats, including: R Data
Serialized
(RDS) and
GeoPackage
(GPKG). The RDS format already contains the time series
corresponding to the samples locations and time period, extracted from
the data cubes used in the RESTORE+ project.
To use these samples in R, install the lulcbrasil-samples package
using devtools:
devtools::install_github("restore-plus/lulcbrasil-samples")Next, load it:
library(lulcbrasilsamples)For other environments, you can rely on the available
GPKG
files. The following steps show the samples metadata and usage examples.
Samples available in this repository contain columns defining their location, time period, and label, encoded as following:
longitude(East-west coordinate in WGS 84).latitude(North-south coordinate in WGS 84).start_date(Initial time series date).end_date(Final time series date).label(Labels associated to the sample).
In the addition to these columns, samples in RDS files contain an
extra column with the time series extracted from the data cubes used in
the Restore+ Project.
time_series(Time series data).
Samples used for training machine leaning models in Restore+ Project.
These samples were built by combining and processing
Base samples.
The following table presents the metadata of this samples:
| Attribute | Details |
|---|---|
| Samples ID | samples_amazon_landsat_2010 |
| Samples region | Amazon biome (Brazil) |
| Samples size | 12512 location points |
| Samples collection methodology | Combining base samples and processing using erosion, self organizing-maps, and visual interpretation. |
| Samples time period | 2010-01-01 to 2010-11-01 |
| Samples labels | Cropland_2_cycles, Forest, Herbaceous_Pasture, Mountainside_Forest, Riparian_Forest, Seasonally_Flooded, Secondary_Vegetation, Semi_Perennial_Crop, Shrubby_Pasture, Silviculture, Wetland |
| Time series source | Extracted from GLAD LANDSAT ARD product available in Open Land Map |
| Time series temporal resolution | 2-month composites (6 data points per year) |
| Time series attributes | BLUE, EVI, GREEN, MNDWI, NBR, NDVI, NIR, RED, SWIR1, SWIR2 |
| Contact point |  Isabel Escada |
| Sponsor | Restore+ Project |
| Hosting institution | National Institute for Space Research - INPE |
| License |  Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) |
To use this sample in R, you can use the following command:
library(lulcbrasilsamples)
data("samples_amazon_landsat_2010")By using the command above, the samples will be available in the
samples_amazon_landsat_2010 variable.
To use this samples in other environment you can download the GPKG
file.
Click to learn how to explore the samples using SITS
If you want to view the samples in R, you can use the sits R package:
library(sits)
plot(samples_amazon_landsat_2010)To view it in am interactive map, use:
sits_view(samples_amazon_landsat_2010)To learn more, please check the sits R package book.
The following table presents the metadata of this samples:
| Attribute | Details |
|---|---|
| Samples ID | samples_amazon_landsat_2022 |
| Samples region | Amazon biome (Brazil) |
| Samples size | 21028 location points |
| Samples collection methodology | Combining base samples and processing using erosion, self organizing-maps, and visual interpretation. |
| Samples time period | 2022-01-01 to 2022-12-01 |
| Samples labels | Cropland_2_cycles,Forest,Herbaceous_Pasture,Mountainside_Forest,Riparian_Forest,Seasonally_Flooded,Secondary_Vegetation,Semi_Perennial_Crop,Shrubby_Pasture,Silviculture,Water,Wetland |
| Time series source | Extracted from LANDSAT-OLI-16D product available in Brazil Data Cube (BDC) |
| Time series temporal resolution | 1-month composites (12 data points per year) |
| Time series attributes | BLUE, EVI, GREEN, MNDWI, NBR, NDVI, NIR08, RED, SWIR16, SWIR22 |
| Contact point | Isabel Escada |
| Sponsor | Restore+ Project |
| Hosting institution | National Institute for Space Research - INPE |
| License | Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) |
To use this sample in R, you can use the following command:
library(lulcbrasilsamples)
data("samples_amazon_landsat_2022")By using the command above, the samples will be available in the
samples_amazon_landsat_2022 variable.
To use this samples in other environment you can download the GPKG
file.
Click to learn how to explore the samples using SITS
If you want to view the samples in R, you can use the sits R package:
library(sits)
plot(samples_amazon_landsat_2022)To view it in am interactive map, use:
sits_view(samples_amazon_landsat_2022)To learn more, please check the sits R package book.
Samples provided by Brazilian agencies, universities, and scientific collaborations. These samples were collected using multiple methodologies, such as field work campaigns and satellite imagery analysis.
The following table presents the metadata of this samples:
| Attribute | Details |
|---|---|
| Samples ID | samples_baixotocantins_landsat_1996 |
| Samples region | Baixo Tocantins in Para state (Brazil) |
| Samples size | 465 location points |
| Samples collection methodology | Visual interpretation using Landsat imagery |
| Samples time period | 1996-01-01 to 1996-12-31 |
| Samples labels | Small-Scale Agriculture - AGPE, All Water Bodies - AGUA, Primary Forest - FLO, Others - OT (Beaches, Outcrops, and Sandbanks), Clean Pasture - PL, Dirty Pasture and Pasture with Regeneration - PSR, Urbanized Area - URB, Advanced Secondary Vegetation - VSA, Initial Secondary Vegetation - VSI |
| Time series source | Extracted from LANDSAT-C2-L2 product available in Microsoft Planetary Computer (MPC) |
| Time series temporal resolution | 3-month composites (4 data points per year) |
| Time series attributes | BLUE, GREEN, RED, NIR08, SWIR16, SWIR22, NDVI, EVI, MNDWI, NBR |
| Contact point | Anielli Rosane de Souza |
| Reference | Souza et al., 2024, Thesis. Available on: “The açaí systems in the amazonian landscape: elements for analysis of the açaí economy in the lower Tocantins region, PA” |
| Hosting institution | National Institute for Space Research - INPE |
| License | Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) |
To use this sample in R, you can use the following command:
library(lulcbrasilsamples)
data("samples_baixotocantins_landsat_1996")By using the command above, the samples will be available in the
samples_baixotocantins_landsat_1996 variable.
To use this samples in other environment you can download the GPKG
file.
Click to learn how to explore the samples using SITS
If you want to view the samples in R, you can use the sits R package:
library(sits)
plot(samples_baixotocantins_landsat_1996)To view it in am interactive map, use:
sits_view(samples_baixotocantins_landsat_1996)To learn more, please check the sits R package book.
The following table presents the metadata of this samples:
| Attribute | Details |
|---|---|
| Samples ID | samples_baixotocantins_landsat_2021 |
| Samples Region | Baixo Tocantins in Para state (Brazil) |
| Samples size | 533 location points |
| Samples collection methodology | Field work campaign |
| Samples time period | 2021-01-01 to 2021-12-31 |
| Samples labels | Small-Scale Agriculture - AGPE, All Water Bodies - AGUA, Primary Forest - FLO, Others - OT (Beaches, Outcrops, and Sandbanks), Clean Pasture - PL, Dirty Pasture and Pasture with Regeneration - PSR, Urbanized Area - URB, Advanced Secondary Vegetation - VSA, Initial Secondary Vegetation - VSI, Large-Scale Agriculture - AGLE2 |
| Time series source | Extracted from LANDSAT-C2-L2 product available in Microsoft Planetary Computer (MPC) |
| Time series temporal resolution | 3-month composites (4 data points per year) |
| Time series attributes | BLUE, GREEN, RED, NIR08, SWIR16, SWIR22,NDVI, EVI, MNDWI, NBR |
| Contact point | Anielli Rosane de Souza |
| Reference | Souza et al., 2024, Thesis. Available on: “The açaí systems in the amazonian landscape: elements for analysis of the açaí economy in the lower Tocantins region, PA” |
| Hosting institution | National Institute for Space Research - INPE |
| License | Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) |
To use this sample in R, you can use the following command:
library(lulcbrasilsamples)
data("samples_baixotocantins_landsat_2021")By using the command above, the samples will be available in the
samples_baixotocantins_landsat_2021 variable.
To use this samples in other environment you can download the GPKG
file.
Click to learn how to explore the samples using SITS
If you want to view the samples in R, you can use the sits R package:
library(sits)
plot(samples_baixotocantins_landsat_2021)To view it in am interactive map, use:
sits_view(samples_baixotocantins_landsat_2021)To learn more, please check the sits R package book.
The following table presents the metadata of this samples:
| Attribute | Details |
|---|---|
| Samples ID | samples_amazon_landsat_2020 |
| Samples region | Amazon Rainforest |
| Samples size | 1489 location points |
| Samples collection methodology | Visual interpretation using Landsat imagery |
| Samples time period | 2020-01-01 to 2020-12-31 |
| Samples labels | Forest |
| Time series source | Extracted from LANDSAT-C2-L2 product available in Microsoft Planetary Computer (MPC) |
| Time series temporal resolution | 3-month composites (4 data points per year) |
| Time series attributes | BLUE, GREEN, RED, NIR08, SWIR16, SWIR22, NDVI, EVI, MNDWI, NBR |
| Contact point | Luis Sadeck |
| Hosting institution | National Institute for Space Research - INPE |
| License | Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) |
To use this sample in R, you can use the following command:
library(lulcbrasilsamples)
data("samples_amazon_landsat_2020")By using the command above, the samples will be available in the
samples_amazon_landsat_2020 variable.
To use this samples in other environment you can download the GPKG
file.
Click to learn how to explore the samples using SITS
If you want to view the samples in R, you can use the sits R package:
library(sits)
plot(samples_amazon_landsat_2020)To view it in am interactive map, use:
sits_view(samples_amazon_landsat_2020)To learn more, please check the sits R package book.
The following table presents the metadata of this samples:
| Attribute | Details |
|---|---|
| Samples ID | samples_rondonia_landsat_1988 |
| Samples Region | Rondonia state (Brazil) |
| Samples size | 1104 location points |
| Samples collection methodology | Visual interpretation using Landsat imagery |
| Samples time period | 1988-01-01 to 1988-12-31 |
| Samples labels | Clear_Cut_Bare_Soil, Clear_Cut_Burned_Area, Clear_Cut_Vegetation, Forest, Water, Moist_Land, Wetland, Moist_Soil, Riparian_Forest, Mountainside_Forest, Eutrophic_Soil, Dystrophic_Soil |
| Time series source | Extracted from LANDSAT-C2-L2 product available in Microsoft Planetary Computer (MPC) |
| Time series temporal resolution | 3-month composites (4 data points per year) |
| Time series attributes | BLUE, GREEN, RED, NIR08, SWIR16, SWIR22, NDVI, EVI, MNDWI, NBR |
| Contact point | Lucas Ferreira |
| Hosting institution | National Institute for Space Research - INPE |
| License | Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) |
To use this sample in R, you can use the following command:
library(lulcbrasilsamples)
data("samples_rondonia_landsat_1988")By using the command above, the samples will be available in the
samples_rondonia_landsat_1988 variable.
To use this samples in other environment you can download the GPKG
file.
Click to learn how to explore the samples using SITS
If you want to view the samples in R, you can use the sits R package:
library(sits)
plot(samples_rondonia_landsat_1988)To view it in am interactive map, use:
sits_view(samples_rondonia_landsat_1988)To learn more, please check the sits R package book.
The following table presents the metadata of this samples:
| Attribute | Details |
|---|---|
| Samples ID | samples_rondonia_landsat_2022 |
| Samples region | Rondonia state (Brazil) |
| Samples size | 6007 location points |
| Samples collection methodology | Visual interpretation using Sentinel imagery |
| Samples time period | 2022-01-01 to 2022-12-31 |
| Samples labels | Clear_Cut_Bare_Soil, Clear_Cut_Burned_Area, Clear_Cut_Vegetation, Forest, Mountainside_Forest, Riparian_Forest, Seasonally_Flooded, Water, Wetland |
| Time series source | Extracted from LANDSAT-C2-L2 product available in Microsoft Planetary Computer (MPC) |
| Time series temporal resolution | 3-month composites (4 data points per year) |
| Time series attributes | BLUE, GREEN, RED, NIR08, SWIR16, SWIR22, NDVI, EVI, MNDWI, NBR |
| Contact point | Gilberto Camara |
| Reference | Souza et at., 2025, Conference. Available On: “An Event-Based Approach For Training Data Selection For Mapping Deforestation” |
| Sponsor | Instituto Clima e Sociedade (iCS) |
| License | Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) |
To use this sample in R, you can use the following command:
library(lulcbrasilsamples)
data("samples_rondonia_landsat_2022")By using the command above, the samples will be available in the
samples_rondonia_landsat_2022 variable.
To use this samples in other environment you can download the GPKG
file.
Click to learn how to explore the samples using SITS
If you want to view the samples in R, you can use the sits R package:
library(sits)
plot(samples_rondonia_landsat_2022)To view it in am interactive map, use:
sits_view(samples_rondonia_landsat_2022)To learn more, please check the sits R package book.
The following table presents the metadata of this samples:
| Attribute | Details |
|---|---|
| Samples ID | samples_aml_landsat_2019 |
| Samples region | Legal Amazon (Brazil) |
| Samples size | 35723 location points |
| Samples collection methodology | Visual interpretation using Sentinel imagery |
| Samples time period | 2019-07-28 to 2020-07-27 |
| Samples labels | agr_semiperene, agr_temp_1, agr_temp_2mais, agric_perene, Corpo_Dagua, Past_Arbu, Past_Herb, silvicultura, Veg_sec |
| Time series source | Extracted from LANDSAT-C2-L2 product available in Microsoft Planetary Computer (MPC) |
| Time series temporal resolution | 3-month composites (4 data points per year) |
| Time series attributes | BLUE, GREEN, RED, NIR08, SWIR16, SWIR22, NDVI, EVI, MNDWI, NBR |
| Contact point | Lidia Bertolo |
| Hosting institution | Brazilian Agricultural Research Corporation - Embrapa |
| License | Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) |
To use this sample in R, you can use the following command:
library(lulcbrasilsamples)
data("samples_aml_landsat_2019")By using the command above, the samples will be available in the
samples_aml_landsat_2019 variable.
To use this samples in other environment you can download the GPKG
file.
Click to learn how to explore the samples using SITS
If you want to view the samples in R, you can use the sits R package:
library(sits)
plot(samples_aml_landsat_2019)To view it in am interactive map, use:
sits_view(samples_aml_landsat_2019)To learn more, please check the sits R package book.
The following table presents the metadata of this samples:
| Attribute | Details |
|---|---|
| Samples ID | samples_aml_landsat_2021 |
| Samples region | Legal Amazon (Brazil) |
| Samples size | 160496 location points |
| Samples collection methodology | Field work campaign |
| Samples time period | 12/07/2021 to 30/09/2022 |
| Samples labels | 1ciclo, 2ciclos, agua, past_arb, past_herb, semiperene, veg_natural |
| Time series source | Extracted from LANDSAT-C2-L2 product available in Microsoft Planetary Computer (MPC) |
| Time series temporal resolution | 3-month composites (4 data points per year) |
| Time series attributes | BLUE, GREEN, RED, NIR08, SWIR16, SWIR22, NDVI, EVI, MNDWI, NBR |
| Contact point | Lidia Bertolo |
| Hosting institution | Brazilian Agricultural Research Corporation - Embrapa |
| License | Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) |
To use this sample in R, you can use the following command:
library(lulcbrasilsamples)
data("samples_aml_landsat_2021")By using the command above, the samples will be available in the
samples_aml_landsat_2021 variable.
To use this samples in other environment you can download the GPKG
file.
Click to learn how to explore the samples using SITS
If you want to view the samples in R, you can use the sits R package:
library(sits)
plot(samples_aml_landsat_2021)To view it in am interactive map, use:
sits_view(samples_aml_landsat_2021)To learn more, please check the sits R package book.