For optimizing production yield while limiting negative environmental impact, sustainable agriculture benefits from real-time, on-the-spot chemical analysis of soil at low cost. Colorimetric paper sensors are ideal candidates, however, their automated readout and analysis in the field is needed. Using mobile technology for paper sensor readout could, in principle, enable the application of machine-learning models for transforming colorimetric data into threshold-based classes that represent chemical concentration. Such a classification method could provide a basis for soil management decisions where high-resolution lab analysis is not required or available. In tropical regions, where reliable soil data is difficult to acquire, this approach would be particularly useful. Here, we report a mobile chemical analysis system based on colorimetric paper sensors that operates under tropical field conditions. A standard smartphone equipped with a dedicated software application automatically classifies the paper sensor results into three classes-low, medium, or high soil pH-which provides a basis for soil correction. The classification task is performed by a machine-learning model which was trained on the colorimetric pH indicators deployed on the paper sensor. By mapping topsoil pH on a test site with an area of 9 hectares, the mobile system was benchmarked in the field against standard soil lab analysis. The mobile system has correctly classified soil pH in 97% of test cases, while reducing the analysis turnaround time from days (soil lab) to minutes (mobile). By performing on-the-spot analyses using the mobile system in the field, a 9-fold increase of spatial resolution reveals pH-variations not detectable in the standard compound mapping mode of lab analysis. We discuss how the mobile analysis can support smallholder farmers and enable sustainable agriculture practices by avoiding excessive soil correction. The system can be extended to perform multi-parameter chemical tests of soil nutrients for applications in environmental monitoring at marginal manufacturing cost.
Copyright: © 2025 Ferreira da Silva et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.