Título:	Satellite imagery to machine learning datasets: an automated system for soil water stress monitoring in agriculture.
Autoria:	HEIDEKER, A. SPERANZA, E. A. FERREIRA, E. J. SILVA, D. KAMIENSKI, C. BIANCHI, R.
Afiliação:	ALEXANDRE HEIDEKER, CENTRO UNIVERSITÁRIO FEI EDUARDO ANTONIO SPERANZA, CNPTIA EDNALDO JOSE FERREIRA, CNPDIA DENER SILVA, CENTRO UNIVERSITÁRIO FEI CARLOS KAMIENSKI REINALDO BIANCHI, CENTRO UNIVERSITÁRIO FEI.
Ano de publicação:	2026
Referência:	In: INTERNATIONAL CONFERENCE ON PRECISION AGRICULTURE, 17., CONGRESSO BRASILEIRO DE AGRICULTURA DE PRECISÃO DIGITAL, 11., 2026, Porto Alegre. Proceedings [...]. Monticello: International Society of Precision Agriculture, 2026.
Páginas:	9 p.
Conteúdo:	Abstract. Satellite remote sensing is a key data source for precision agriculture, enabling vegetation indices such as NDMI and EVI to monitor vegetation health and soil water stress over large areas. Its practical use, however, is often constrained by manual image selection, downloading, and preprocessing workflows, which are time-consuming, hard to reproduce, and ill-suited to building the large, multi-temporal datasets that machine learning requires. This study presents an automated system for satellite image acquisition and preprocessing that enables scalable dataset generation for soil water stress monitoring. The system retrieves historical and current Sentinel-2 multispectral imagery through the Copernicus API, guided by predefined spatial and temporal parameters. Areas of interest are defined as multiple polygons in GeoJSON files representing individual agricultural sectors, which are grouped within a simplified bounding area to reduce redundant queries and improve retrieval efficiency. By eliminating manual scene inspection, the system reduces human effort while ensuring consistency and reproducibility, enabling systematic expansion of monitored regions and supporting the creation of high-quality datasets for training and deploying machine learning models in irrigation management. Its modular design accommodates the integration of additional satellite missions, spectral indices, and preprocessing methods in the future
Thesagro:	Sensoriamento Remoto Agricultura de Precisão
NAL Thesaurus:	Remote sensing Vegetation index Water stress Crop management
Palavras-chave:	Sentinel-2 Índice de vegetação Aprendizado de máquina Estresse hídrico do solo Monitoramento da cultura do café Satellite remote sensing Vegetation indices Machine learning datasets Cloud masking Soil water stress Coffee crop monitoring
Notas:	ICPA 2026, ConBAP 2026. Na publicação: Eduardo Speranza, Ednaldo Pereira.
Tipo do material:	Artigo em anais e proceedings
Acesso:	openAccess
Aparece nas coleções:	Artigo em anais de congresso (CNPTIA)