Satellite-ground linking for agronomic monitoring using the PlantAR application

Abstract

PlantAR is an open-source smartphone application designed to support georeferenced agronomic monitoring through the acquisition of ground-level images linked to points of interest detected via satellite imagery. The application allows manual or automated image capture, supported by temporal, geographic, and sensor metadata (accelerometer and magnetometer). It provides real-time mapping and audio alerts that trigger automatic image capture when entering a predefined radius. As an initial validation, the study focused on automating the acquisition of georeferenced ground-level images near roads, linked to satellite-based NDVI time series. PlantAR guided the acquisition based on coordinates and areas of interest loaded in GeoJSON format. Ground-level images were collected along designated routes and compared to the satellite-derived NDVI curve. The results show a strong temporal correspondence between the evolution of crop growth stages and the NDVI signal, supported by a statistical correlation with RGB indices (notably the Green Chromatic Coordinate, GCC) calculated from the terrestrial images. PlantAR proved to be practical and adaptable, offering flexibility to capture images from the road or in closer proximity to the crop. The application provides a scalable solution for visual field monitoring and serves as a foundation for future developments in satellite–ground data integration.