| |
|
|
|
Registro Completo |
|
Biblioteca(s): |
Embrapa Solos. |
|
Data corrente: |
08/02/2022 |
|
Data da última atualização: |
11/03/2022 |
|
Tipo da produção científica: |
Artigo em Anais de Congresso |
|
Autoria: |
KUCHLER, P. C.; SIMÕES, M.; BEGUE, A.; FERRAZ, R. P. D. |
|
Afiliação: |
PATRICK CALVANO KUCHLER, UERJ; MARGARETH GONCALVES SIMOES, CNPS; AGNÈS BEGUE, CIRAD; RODRIGO PECANHA DEMONTE FERRAZ, CNPS. |
|
Título: |
Big earth observation data and machine learning for mapping crop-livestock integrated system in Brazil. |
|
Ano de publicação: |
2021 |
|
Fonte/Imprenta: |
In: WORLD CONGRESS ON INTEGRATED CROP-LIVESTOCK-FORESTRY SYSTEMS, 2., 2021. WCCLF 2021 proceedings. Brasília, DF: Embrapa, 2021. p. 904-909. WCCLF 2021. Evento online. |
|
Idioma: |
Inglês |
|
Conteúdo: |
The adoption of crop-livestock (iCL) integrated systems has been pointed out as an important strategy for increasing production based on sustainable intensification of land use in Brazil. Mapping and monitoring the iCL areas would allow us to know the expansion rates and the adoption level of the integrated system, being an important instrument for public policy management. However, due to the time-space variability from integrated production systems, developing methods based on remote sensing remains a major challenge. In this sense, this work discusses the application of Big Data and machine learning concepts in Earth Observation Data as a strategy to compose a methodology for monitoring the iCL in Brazil. We tested the capacity of the Random Forest (RF) classifier applied to MODIS time series to iCL detection in the Mato Grosso State, Brazil. For this, we evaluated the classification accuracy for the years between 2012 and 2019, totaling 3,864 images processed. The overall accuracy founded was between 0.77 and 0.89 and an fscore average of 0.85 was found for the iCL class. The generated maps showed a trajectory of sustainable intensification, with the expansion of the iCL area from 1,100,000 ha in 2012/2013 to 2,597,000 ha in 2018/2019, an increase of 135%. The results indicate that the use of the RF classification technique with MODIS times series has great potential to compose an iCL monitoring methodology, requiring parallel and cloud computing applied to advanced algorithms. MenosThe adoption of crop-livestock (iCL) integrated systems has been pointed out as an important strategy for increasing production based on sustainable intensification of land use in Brazil. Mapping and monitoring the iCL areas would allow us to know the expansion rates and the adoption level of the integrated system, being an important instrument for public policy management. However, due to the time-space variability from integrated production systems, developing methods based on remote sensing remains a major challenge. In this sense, this work discusses the application of Big Data and machine learning concepts in Earth Observation Data as a strategy to compose a methodology for monitoring the iCL in Brazil. We tested the capacity of the Random Forest (RF) classifier applied to MODIS time series to iCL detection in the Mato Grosso State, Brazil. For this, we evaluated the classification accuracy for the years between 2012 and 2019, totaling 3,864 images processed. The overall accuracy founded was between 0.77 and 0.89 and an fscore average of 0.85 was found for the iCL class. The generated maps showed a trajectory of sustainable intensification, with the expansion of the iCL area from 1,100,000 ha in 2012/2013 to 2,597,000 ha in 2018/2019, an increase of 135%. The results indicate that the use of the RF classification technique with MODIS times series has great potential to compose an iCL monitoring methodology, requiring parallel and cloud computing applied to advanced algo... Mostrar Tudo |
|
Palavras-Chave: |
Machine learning; MODIS time series. |
|
Thesagro: |
Agricultura Sustentável. |
|
Thesaurus Nal: |
Sustainable agricultural intensification. |
|
Categoria do assunto: |
P Recursos Naturais, Ciências Ambientais e da Terra |
|
URL: |
https://ainfo.cnptia.embrapa.br/digital/bitstream/item/231050/1/Big-earth-observation-data-and-machine-learning-2021.pdf
|
|
Marc: |
LEADER 02262nam a2200193 a 4500
001 2139789
005 2022-03-11
008 2021 bl uuuu u00u1 u #d
100 1 $aKUCHLER, P. C.
245 $aBig earth observation data and machine learning for mapping crop-livestock integrated system in Brazil.$h[electronic resource]
260 $aIn: WORLD CONGRESS ON INTEGRATED CROP-LIVESTOCK-FORESTRY SYSTEMS, 2., 2021. WCCLF 2021 proceedings. Brasília, DF: Embrapa, 2021. p. 904-909. WCCLF 2021. Evento online.$c2021
520 $aThe adoption of crop-livestock (iCL) integrated systems has been pointed out as an important strategy for increasing production based on sustainable intensification of land use in Brazil. Mapping and monitoring the iCL areas would allow us to know the expansion rates and the adoption level of the integrated system, being an important instrument for public policy management. However, due to the time-space variability from integrated production systems, developing methods based on remote sensing remains a major challenge. In this sense, this work discusses the application of Big Data and machine learning concepts in Earth Observation Data as a strategy to compose a methodology for monitoring the iCL in Brazil. We tested the capacity of the Random Forest (RF) classifier applied to MODIS time series to iCL detection in the Mato Grosso State, Brazil. For this, we evaluated the classification accuracy for the years between 2012 and 2019, totaling 3,864 images processed. The overall accuracy founded was between 0.77 and 0.89 and an fscore average of 0.85 was found for the iCL class. The generated maps showed a trajectory of sustainable intensification, with the expansion of the iCL area from 1,100,000 ha in 2012/2013 to 2,597,000 ha in 2018/2019, an increase of 135%. The results indicate that the use of the RF classification technique with MODIS times series has great potential to compose an iCL monitoring methodology, requiring parallel and cloud computing applied to advanced algorithms.
650 $aSustainable agricultural intensification
650 $aAgricultura Sustentável
653 $aMachine learning
653 $aMODIS time series
700 1 $aSIMÕES, M.
700 1 $aBEGUE, A.
700 1 $aFERRAZ, R. P. D.
Download
Esconder MarcMostrar Marc Completo |
|
Registro original: |
Embrapa Solos (CNPS) |
|
|
Biblioteca |
ID |
Origem |
Tipo/Formato |
Classificação |
Cutter |
Registro |
Volume |
Status |
URL |
Voltar
|
|
|
| Registros recuperados : 1 | |
| Registros recuperados : 1 | |
|
| Expressão de busca inválida. Verifique!!! |
|
|
