Determination of Moisture Content in Concrete Aggregates using Machine Learning algorithms and Hyperspectral Imaging

The quantification of moisture content in various economic areas and in different industrial processes has been a parameter investigated over many years because it serves to estimate the quality, durability and other important parameters at commercial and environmental level. This paper presents a summary of the advancement obtained in recent years in moisture measurement techniques, as well as a new classification of the most representative methods, as mentioned in research and scientific articles. The applications of traditional direct techniques, such as the Karl Fischer titration or the thermogravimetric method are discussed, as well as approaches that use NIR image processing, neural networks, or microwaves, among others. Environmental applications such as soil moisture measurement using radiometry and prediction algorithms are reviewed as well. Furthermore, the most prominent methods are analysed in detail, describing the way they are performed, their advantages and disadvantages, the most relevant applications and the main challenges that should be investigated further.