Development of a software based on automatic multi-temporal aerial images classification to assess retrospective environmental exposures to pesticides in epidemiological studies

Élodie Faure (1) , Rémi Ratajczak (2, 3, 1) , Carlos F Crispim-Junior (3) , Olivia Pérol (1) , Laure Tougne (3) , Béatrice Fervers (1)

1 Département cancer environnement (Centre Léon Bérard - Lyon)
2 ADEME - Agence de l'Environnement et de la Maîtrise de l'Energie
3 imagine - Extraction de Caractéristiques et Identification

Environmental exposure to agricultural pesticides (EEAP) resulting from the drift of agricultural pesticides applied on fields and is suspected to be a risk factor for several diseases. Geographic Information Systems (GIS) are increasingly used in environmental epidemiological studies to assess EEAP. Crop acreage proximate to subjects residences has been suggested as a surrogate for EEAP. Retrospective characterization of EEAP is essential due to long latency. Earlier data are lacking and limitting the capacity to capture the life-course effects of exposure. We aimed to develop an innovative software to analyze the historical monochromatic aerial images to reconstruct land use to characterize EEAP retrospectively.

Domaines: Apprentissage [cs.LG]; Vision par ordinateur et reconnaissance de formes [cs.CV]; Génie logiciel [cs.SE]; Géographie; Santé; Environnement et Société

Élodie Faure, Rémi Ratajczak, Carlos F Crispim-Junior, Olivia Pérol, Laure Tougne, et al.. Development of a software based on automatic multi-temporal aerial images classification to assess retrospective environmental exposures to pesticides in epidemiological studies. CLARA 2018 Cancer Research Forum, Apr 2018, Lyon, France. ⟨hal-01784654⟩