SOIL ORGANIC MATTER DETERMINATION USING WIRELESS SENSOR NETWORKS

Abstract

   The paper addresses the problem of predicting soil organic matter content in an agricultural field using information collected by a low-cost network of mobile, wireless and noisy sensors that can take discrete measurements in the environment. In this context, it is proposed that the spatial phenomenon of organic matter in soil to be monitored is modeled using Gaussian processes. The proposed model then enables the wireless sensor network to estimate the soil organic matter at all unobserved locations of interest. The estimated values at predicted locations are highly comparable to those at corresponding points on a realistic image that is aerially taken by a very expensive and complex remote sensing system.