ABSTRACT
Our approach to multispectral remote sensing assessment of constructive pathologies has been organized in two different parts. The first one (Chapter 7) dealt with equipment and methods. The second one, developed in this chapter, will be related to the application to real cases, encompassing relevant aspects such as: data acquisition (sensor type choice and field works planning), processing (filtering and segmentation), sensor registration, true orthophoto generation, orthophoto classifications (through supervised and unsupervised techniques) and the qualitative and quantitative analysis of results (by means of confusion matrix, spectral separability, overall accuracy, reliability and agreement of informational classes). Based on these premises three case studies have been addressed. On the one hand, two historical stone masonry constructions and on the other hand one modern reinforced concrete construction. These three case studies will be used as examples of best practices in multispectral dataset management and processing, and will serve to evaluate the flexibility of the methodology proposed for detecting and classifying accurately a wide range of constructive pathologies.
