Enhancement techniques for abnormality detection using thermal image

Abstract

Digital thermal imaging is considered as a non-invasive diagnostic tool and a real-time monitoring technique for indicating the physiological changes of the underlying tissue from the superficial thermal signature. A thermal camera can detect temperature variations in the body, as low as 0.1°C. The observed colour pattern depends on the prevailing temperature of the target in a controlled environment. This colour-based thermal pattern is further processed for identifying abnormalities. This process of identification is done by applying various methods such as histogram equalisation, Otsu thresholding and morphologic function. These steps are applied to thermal images of a foot acquired from volunteers and abnormalities were identified. The identification process was based on a threshold obtained from the histogram and it was found to be in the range of 76–80.