Mapping deforested areas in the Cerrado Biome through recurrent neural networks

B. M. Matosak; R. V. Maretto; T. S. Korting; M. Adami; L. M.G. Fonseca

doi:10.1109/IGARSS39084.2020.9324019

Mapping deforested areas in the Cerrado Biome through recurrent neural networks

B. M. Matosak, R. V. Maretto, T. S. Korting, M. Adami, L. M.G. Fonseca

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

1 Citation (Scopus)

Abstract

The Brazilian Savannah, also known as Cerrado Biome, is a hotspot for the Brazilian biodiversity and is also important for this country water supply. One of the most active Brazilian agricultural frontiers, the region has a history of primary vegetation suppression. Accurately map this phenomenon is an important step to inform and enable government conservation programs. In this work, we used a Long Short-Term Memory network to generate a deforestation map for the Cerrado. The PRODES deforestation inventory was used as ground truth during training and evaluation. We used as inputs a dense Landsat 8 time series composed by 6 spectral bands and 3 vegetation indices, as well as the SRTM terrain slope. The methodology was tested on an area comprising about 31,450 km², achieving approximately 98.5% global accuracy.

Original language	English
Title of host publication	2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020 - Proceedings
Publisher	IEEE
Pages	1389-1392
Number of pages	4
ISBN (Electronic)	9781728163741
DOIs	https://doi.org/10.1109/IGARSS39084.2020.9324019
Publication status	Published - 26 Sep 2020
Externally published	Yes
Event	2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020 - Virtual, Waikoloa, United States Duration: 26 Sep 2020 → 2 Oct 2020

Conference

Conference	2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020
Abbreviated title	IGARSS 2020
Country/Territory	United States
City	Virtual, Waikoloa
Period	26/09/20 → 2/10/20

Keywords

Brazilian Savannah
Cerrado Biome
Deep Learning
Deforestation
LSTM
ITC-CV
n/a OA procedure

Access to Document

10.1109/IGARSS39084.2020.9324019

Cite this

@inproceedings{9a5060e4c986477b801d437ddacbdeaa,

title = "Mapping deforested areas in the Cerrado Biome through recurrent neural networks",

abstract = "The Brazilian Savannah, also known as Cerrado Biome, is a hotspot for the Brazilian biodiversity and is also important for this country water supply. One of the most active Brazilian agricultural frontiers, the region has a history of primary vegetation suppression. Accurately map this phenomenon is an important step to inform and enable government conservation programs. In this work, we used a Long Short-Term Memory network to generate a deforestation map for the Cerrado. The PRODES deforestation inventory was used as ground truth during training and evaluation. We used as inputs a dense Landsat 8 time series composed by 6 spectral bands and 3 vegetation indices, as well as the SRTM terrain slope. The methodology was tested on an area comprising about 31,450 km2, achieving approximately 98.5% global accuracy.",

keywords = "Brazilian Savannah, Cerrado Biome, Deep Learning, Deforestation, LSTM, ITC-CV, n/a OA procedure",

author = "Matosak, {B. M.} and Maretto, {R. V.} and Korting, {T. S.} and M. Adami and Fonseca, {L. M.G.}",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020, IGARSS 2020 ; Conference date: 26-09-2020 Through 02-10-2020",

year = "2020",

month = sep,

day = "26",

doi = "10.1109/IGARSS39084.2020.9324019",

language = "English",

pages = "1389--1392",

booktitle = "2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020 - Proceedings",

publisher = "IEEE",

}

Matosak, BM, Maretto, RV, Korting, TS, Adami, M & Fonseca, LMG 2020, Mapping deforested areas in the Cerrado Biome through recurrent neural networks. in 2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020 - Proceedings., 9324019, IEEE, pp. 1389-1392, 2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020, Virtual, Waikoloa, United States, 26/09/20. https://doi.org/10.1109/IGARSS39084.2020.9324019

Mapping deforested areas in the Cerrado Biome through recurrent neural networks. / Matosak, B. M.; Maretto, R. V.; Korting, T. S. et al.

2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020 - Proceedings. IEEE, 2020. p. 1389-1392 9324019.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Mapping deforested areas in the Cerrado Biome through recurrent neural networks

AU - Matosak, B. M.

AU - Maretto, R. V.

AU - Korting, T. S.

AU - Adami, M.

AU - Fonseca, L. M.G.

PY - 2020/9/26

Y1 - 2020/9/26

N2 - The Brazilian Savannah, also known as Cerrado Biome, is a hotspot for the Brazilian biodiversity and is also important for this country water supply. One of the most active Brazilian agricultural frontiers, the region has a history of primary vegetation suppression. Accurately map this phenomenon is an important step to inform and enable government conservation programs. In this work, we used a Long Short-Term Memory network to generate a deforestation map for the Cerrado. The PRODES deforestation inventory was used as ground truth during training and evaluation. We used as inputs a dense Landsat 8 time series composed by 6 spectral bands and 3 vegetation indices, as well as the SRTM terrain slope. The methodology was tested on an area comprising about 31,450 km2, achieving approximately 98.5% global accuracy.

AB - The Brazilian Savannah, also known as Cerrado Biome, is a hotspot for the Brazilian biodiversity and is also important for this country water supply. One of the most active Brazilian agricultural frontiers, the region has a history of primary vegetation suppression. Accurately map this phenomenon is an important step to inform and enable government conservation programs. In this work, we used a Long Short-Term Memory network to generate a deforestation map for the Cerrado. The PRODES deforestation inventory was used as ground truth during training and evaluation. We used as inputs a dense Landsat 8 time series composed by 6 spectral bands and 3 vegetation indices, as well as the SRTM terrain slope. The methodology was tested on an area comprising about 31,450 km2, achieving approximately 98.5% global accuracy.

KW - Brazilian Savannah

KW - Cerrado Biome

KW - Deep Learning

KW - Deforestation

KW - LSTM

KW - ITC-CV

KW - n/a OA procedure

UR - https://ezproxy2.utwente.nl/login?url=https://doi.org/10.1109/IGARSS39084.2020.9324019

U2 - 10.1109/IGARSS39084.2020.9324019

DO - 10.1109/IGARSS39084.2020.9324019

M3 - Conference contribution

AN - SCOPUS:85101999738

SP - 1389

EP - 1392

BT - 2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020 - Proceedings

PB - IEEE

T2 - 2020 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2020

Y2 - 26 September 2020 through 2 October 2020

ER -

Mapping deforested areas in the Cerrado Biome through recurrent neural networks

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this