Algoritmo de segmentación de regiones en paredes mediante análisis temporal de mapas térmicos

Abstract

Nowadays, moistures and thermal leaks in buildings are manually detected by an operator, who roughly delimits those critical regions in thermal images. Nevertheless, the use of AI techniques can greatly improve the manual thermal analysis, providing automatically more precise and objective results. This paper presents a temporal-clustering based technique that carries out the segmentation of thermal orthoimages (TO) of a wall, which have been taken at different times. The algorithm has two stages: region labelling and consensus. In order to delimit regions with similar temperature, a clustering procedure is applied to each TO, obtaining a set of labelled TOs. As a result, a three-dimensional data structure XYt is obtained. Dimensions XY correspond to a labelled TO and dimension t is the time of the session. In the second stage, a consensus algorithm between corresponding regions at different times is applied. Thus, the method delimitates regions with different thermal behaviour over time, which are characterized with a set of statistical indicators. The approach has been tested in real scenes by using a 3D thermal scanner. A case study, composed of 48 thermal orthoimages at 30 minute intervals over 24 hours, are presented.