A low-cost IoT solution for predicting foliar diseases in agroecological crops at a family farming scale

Experience in the peri-urban area of La Plata

Abstract

This work is framed within two projects: the first, an R&D&I project titled Automatic prediction of leaf diseases in leaf crops produced in greenhouses under agroecological management”, developed by the Research Laboratory on New Information Technologies at the National University of La Plata (LINTI-UNLP); and the second, a Technological Linkage project of UNLP. Its objective is to contribute to the development of low-cost digital technologies for agroecological production at the scale of family farming, facilitating the preventive management of foliar diseases caused by fungi and pseudofungi. For this purpose, automatic sensing of microenvironmental conditions in greenhouses is employed using IoT technologies. The project is being carried out in the horticultural area of peri-urban La Plata, specifically in two farms belonging to producer families in the locality of Arana. In this context, this paper analyzes: a) the learnings from the 'ad hoc' IoT solution deployed, and b) the data collection during a 5-and-a-half-month experimental stage. The latter point evaluates the amount of data received, the possible causes of losses, the quality of the data, and the identification of environmental thresholds relevant to the problem under study. The initial findings of this analysis allow us to determine the magnitude of losses and their relevance to the problem, identify anomalous data requiring normalization, and detect discrepancies in measurements obtained from different devices. These aspects are fundamental for the development of a predictive model for the detection of leaf diseases, which is the main objective of these projects.