Título:	Agricultural property monitoring using geospatial foundation models.
Autoria:	GOULART, A. PAPA, M. BERGIER, I. DOURADO, D.
Afiliação:	ANTONIO GOULART, UNIVERSIDADE DE SÃO PAULO; MATHEUS PAPA; IVAN BERGIER TAVARES DE LIMA, CNPTIA; DURVAL DOURADO, UNIVERSIDADE DE SÃO PAULO.
Ano de publicação:	2025
Referência:	In: WORKSHOP CIENTÍFICO DO CENTRO DE CIÊNCIA PARA O DESENVOLVIMENTO EM AGRICULTURA DIGITAL – SEMEAR DIGITAL, 2., 2025, Campinas. Anais [...]. Piracicaba: ESALQ/USP, 2025. p. 33-40.
Conteúdo:	In this paper we present a preliminary study on the application of geospatial foundation models to the satellite-based monitoring of agricultural properties. Images from the Sentinel-2 mission are processed with the Clay foundation model in order to produce embeddings in the model’s latent space, which are later reduced in dimension with a principal component analysis and plotted in 2 dimensions to enable visual inspection. A case study was carried out considering scenes from a property located in the Guia Lopes da Laguna agrotechnological district, a partner of Semear Digital. Two management situations were taken into account: a) traditional pasture; and b) initial preparation for an integrated livestock-forest system. Differences in the embeddings from both scenarios were analyzed, and early observations indicate that the method has potential and may become a consolidated practice in the area.
Thesagro:	Sensoriamento Remoto
NAL Thesaurus:	Remote sensing
Palavras-chave:	Aprendizado de máquina Aprendizado profundo Integração pecuária-floresta Projeto Semear Digital Deep machine learning Livestock-forest integration
ISBN:	978-85-86481-94-9
Notas:	Organização: Silvia Maria Fonseca Silveira Massruhá, Durval Dourado Neto, Luciana Alvim Santos Romani, Jayme Garcia Arnal Barbedo, Édson Luis Bolfe, Ivan Bergier, Maria Angelica de Andrade Leite, Vitor Del Alamo Guarda, Catarina Barbosa Careta.
Tipo do material:	Artigo em anais e proceedings
Acesso:	openAccess
Aparece nas coleções:	Artigo em anais de congresso (CNPTIA)