Accessible water quality monitoring through hybrid human–machine colorimetric methods

Effective water quality monitoring is important for environmental protection and public health, yet conventional field and laboratory methods each present significant limitations. Field tools such as colorimetric test strips offer affordability and accessibility but are prone to subjective interpretation and environmental variability. In contrast, laboratory-based techniques provide high precision but are costly, resource-intensive, and less feasible in decentralized contexts. This study presents a hybrid human-machine methodology that improves the accuracy and reproducibility of colorimetric test strip analysis while maintaining field-level accessibility. A total of 34 water samples collected along a 7-km stretch of Seunggi Stream in Incheon, South Korea, were analyzed using a web-based platform that extracts RGB values from images of test strips and reference charts. To translate color into concentration, the system calculates Euclidean distances between test strip colors and known reference values, then applies inverse distance weighting (IDW) to interpolate continuous estimates from the closest matches. This approach overcomes the limitations of discrete reference charts, enabling more precise and reproducible readings without the need for complex machine learning models. Validation against standard laboratory methods revealed strong correlations (r > 0.85 for pH, lead, and total hardness), supporting the reliability of the approach. Spatial trends in pollutants were successfully mapped, demonstrating the method's utility for environmental monitoring. This cost-effective, scalable solution bridges the gap between subjective field testing and laboratory precision, offering a practical tool for resource-limited settings, citizen science, and preliminary assessments. Future research will refine analyte-specific accuracy and expand applicability to more diverse conditions.