Enhancement of infrared image for roof leakage detection

Abstract

Roof Leakage is the major problem faced in almost all buildings irrespective of the age of the building. Basically, roof leak is caused because of the different conditions and material used in construction. But most of the time these leakages are so small that we cannot see with bare eyes. That's why infrared cameras are used for detecting roof leaks. Based on the images captured by infrared cameras, edge map is drawn and that particular area is repaired. Because of the thermal nature of the infrared image, we cannot see edges perfectly which makes us to use approximation of an edge. This leads to the wastage of the material used in repairing the roof leaks which in turn increases the cost of repair also. This paper presents a technique of infrared image enhancement for roof leakage detection. This technique combines the benefits of homomorphic image processing and the additive wavelet transform. The idea behind this technique is to decompose the image into three sub bands in an additive fashion using an additive wavelet transform. The homomorphic processing is performed on each sub band, separately. The homomorphic processing transforms the image into illumination and reflectance components. Butterworth filter is used to enhance the reflection and illumination components in the image, separately. Finally, an inverse additive wavelet transform is performed on the homomorphic enhanced sub bands to get an infrared image with better visual details as well as edge map.