A knowledge-based system for the non-destructive diagnostics of façade isolation using the information fusion of visual and IR images

Abstract

This paper describes a knowledge-based system for the non-destructive diagnostics of façade isolation. The system uses the fusion of information extracted from images obtained from two electromagnetic wavebands: from low-resolution IR images (320 × 240 pixels; 7.5–13 μm wavelength), and from high-resolution visual RGB images (3264 × 2448 pixels; 0.39–0.75 μm wavelength). The fusion of the information is a perquisite for the design of an efficient knowledge-based system for non-destructive testing and façade diagnostics. In the paper, special attention is devoted to the following knowledge-based system components: system-core components (the database, the knowledge base, the inference engine, and the user interface) and application-specific components (image acquisition, the resolution pyramid, image pre-processing and feature extraction, image registration, object segmentation, fusion, and data analysis). Examples of the non-destructive testing and the façade diagnostics obtained with the system are presented.