Evaluating the Potential of Multispectral Sensor Networks based on Deployments in Wheat Fields

The reflection spectra of vegetation hold valuable information on the state of the plants. Analysing data from satellites, airplanes, and Unmanned Aerial Vehicles (UAVs) equipped with multispectral sensors has therefore become very popular in the last decades. While these systems offer global coverage in the case of satellites and high spatial resolution in the case of UAVs, these systems are not suitable for obtaining continuous, real-time measurements. With advances in miniaturization, Wireless Sensor Networks (WSNs) are on the brink of becoming capable to serve this purpose. While different kinds of sensors have been experimentally evaluated for agricultural WSNs, this has barely been done with multispectral sensors. Therefore, we share our results from two deployments of such WSNs. We provide comparisons with other sensing system, investigate the suitability for generating Normalized Difference Vegetation Index (NDVI) timeseries, and sneak a peek at the derivation of yield estimates. Index Terms—Multispectral Imaging, Wireless Sensor Networks, Precision Agriculture