Applications, challenges and perspectives for monitoring agricultural dynamics in the Brazilian savanna with multispectral remote sensing

Autor(es)

Parreiras, Taya Cristo ; Bolfe, Edson Luis ; Pereira, Paulo Roberto Mendes ; de Souza, Abner Matheus ; Alves, Vinicius Fernandes

• Tipo de documento: Artigo Científico
• Fonte: REMOTE SENSING APPLICATIONS-SOCIETY AND ENVIRONMENT; v. 37, p. 18-pg., 2025-01-21.

Resumo

Land use and cover changes significantly impact landscape configuration, climate change, and society. The processes of expansion, conversion, intensification, diversification, and reduction materialize these changes in the agricultural environment. The Cerrado, or Brazilian Savanna, is a biodiversity hotspot, extremely important for water production, and one of the most important biomes for global food production. In this sense, monitoring agricultural dynamics in this environment plays a crucial role in sustainable planning, assessment of environmental impacts, and food security. In this study, we propose to analyze the evolution of the role of multispectral orbital remote sensing in mapping and monitoring agricultural dynamics processes in the Cerrado. Therefore, a narrative review of the literature based on studies developed in the biome was carried out to identify advances in tools, processes, and resources, as well as evaluate the challenges and perspectives for the future. Among other relevant results, monitoring these processes has become faster, more frequent, and more accurate, mainly through the combined use of high temporal resolution time series of spectral data and machine learning algorithms. Promising results have been obtained with Harmonized Landsat Sentinel-2 (HLS) data. The consolidation of deep neural networks has contributed substantially to detecting and delimitating complex intensification and diversification systems, such as central irrigation pivots and intercropping. However, there are challenges and obstacles to be faced, such as expanding the use of Sentinel-2 data, establishing means for sharing field data, and expanding studies to more fragmented landscapes, especially agricultural production on small properties. (AU)

• Processo FAPESP: 24/13150-5 - Análise espaço-temporal da dinâmica agrícola em região de alta diversidade produtiva partir de imagens multisensor e aprendizado de máquina
• Beneficiário: Taya Cristo Parreiras
• Modalidade de apoio: Bolsas no Brasil - Doutorado
• Processo FAPESP: 22/09319-9 - Centro de Ciência para o Desenvolvimento em Agricultura Digital - CCD-AD/SemeAr
• Beneficiário: Silvia Maria Fonseca Silveira Massruhá
• Modalidade de apoio: Auxílio à Pesquisa - Centros de Ciência para o Desenvolvimento