Non-Invasive Moisture Sensing and Classification Using Ultra-Wideband Signals

This paper presents a novel method for sensing and classifying water content in dielectric materials. Moisture levels are obtained by non-invasive methods using digitally processed electromagnetic signals. Those signals are the scattering produced within the measured material when illuminated with an UltraWideband pulse. The pulse generation, signal detection, and discretization hardware was built on a custom design developed at INTI. The collected information is processed through a non-linear pipeline to finally perform classification according to the water content of the target. We show that the precision of the proposed sensing architecture is comparable with standard methods thatmay be invasive. As a non-invasive method, this new sensor has great potential for applications such as farming and construction industries, among others.
Fil: Albert, Raymundo Jose. Instituto Nacional de Tecnología Industrial; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina
Fil: Marchi, Edgardo Jose. Instituto Nacional de Tecnología Industrial; Argentina