Detection of Residual Salinity on Road Surface Using Neural Network

This article discusses the use of microwaves to detect saline solutions on road surfaces in winter. The reflection characteristics of water and saline solutions at microwave frequencies depend on the respective dielectric constants. Therefore, the salinity and thickness of films seem to be detectable based on reflection characteristics. We confirm this ability through theoretical calculations and verification experiments. Field tests verify the practicality of this approach. Furthermore, we propose a new estimation method for the relationships salinity level and reflection characteristics using neural networks. It is shown that the proposed method has a good performance.