Infrared Thermography (IRT) is the widely used Non-Destructive Testing (NDT) technique for health monitoring of buildings. IRT uses an IR camera that captures the temperature variations and maps it into thermographs. Under normal conditions, temperature distribution is uniform. On the other hand an abnormality appears either as hot or cold spot. Hence interpretation of these thermographs provides information about the abnormality. Various image segmentation techniques are cited in literature for abnormality detection in visible images. However an efficient domain specific segmentation algorithm for thermographs is yet to be evolved. Local Intensities Operation (LIO) based segmentation is already proposed for thermal image segmentation. However, it has failed miserably for low contrast thermographs. Hence an enhanced approach is proposed in this paper that performs contrast stretching based LIO for segmentation. From subjective analysis, it is found that the proposed technique provides better results even for low contrast thermal images.
SANTHOSH, R.R.; NANDHITHA, N. M.; and SHEELARANI, B.
"NOVEL IMAGE SEGMENTATION TECHNIQUE FOR LOW CONTRAST BUILDING THERMOGRAPHS,"
International Journal of Computer and Communication Technology: Vol. 7:
1, Article 18.
Available at: https://www.interscience.in/ijcct/vol7/iss1/18