%0 Journal Article
%4 sid.inpe.br/mtc-m21d/2023/01.27.12.28
%2 sid.inpe.br/mtc-m21d/2023/01.27.12.28.32
%@doi 10.3390/rs15020521
%@issn 2072-4292
%T Mapping Tropical Forest Cover and Deforestation with Planet NICFI Satellite Images and Deep Learning in Mato Grosso State (Brazil) from 2015 to 2021
%D 2023
%8 Jan.
%9 journal article
%A Wagner, Fabien Hubert,
%A Silva, Ricardo Dalagnol da,
%A Silva-Junior, Celso Henrique Leite,
%A Carter, Griffina,
%A Ritz, Alison L.,
%A Hirye, Mayumi C. M.,
%A Ometto, Jean Pierre Henry Balbaud,
%A Saatchi, Sassan,
%@affiliation University of California
%@affiliation University of California
%@affiliation University of California
%@affiliation University of California
%@affiliation CTREES, Pasadena
%@affiliation Universidade de São Paulo (USP)
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation University of California
%@electronicmailaddress fhwagner@ucla.edu
%@electronicmailaddress ricds@hotmail.com
%@electronicmailaddress
%@electronicmailaddress
%@electronicmailaddress
%@electronicmailaddress
%@electronicmailaddress jean.ometto@inpe.br
%B Remote Sensing
%V 15
%N 2
%P e521
%K land-cover and land-use, semantic segmentation, TensorFlow 2, tropical forests, U-net.
%X Monitoring changes in tree cover for assessment of deforestation is a premise for policies to reduce carbon emission in the tropics. Here, a U-net deep learning model was used to map monthly tropical tree cover in the Brazilian state of Mato Grosso between 2015 and 2021 using 5 m spatial resolution Planet NICFI satellite images. The accuracy of the tree cover model was extremely high, with an F1-score >0.98, further confirmed by an independent LiDAR validation showing that 95% of tree cover pixels had a height >5 m while 98% of non-tree cover pixels had a height <5 m. The biannual map of deforestation was then built from the monthly tree cover map. The deforestation map showed relatively consistent agreement with the official deforestation map from Brazil (67.2%) but deviated significantly from Global Forest Change (GFC)s year of forest loss, showing that our product is closest to the product made by visual interpretation. Finally, we estimated that 14.8% of Mato Grossos total area had undergone clear-cut logging between 2015 and 2021, and that deforestation was increasing, with December 2021, the last date, being the highest. High-resolution imagery from Planet NICFI in conjunction with deep learning techniques can significantly improve the mapping of deforestation extent in tropical regions.
%@language en
%3 remotesensing-15-00521.pdf