Infrared Thermal Imaging for Intelligent Leakage Detection in Underground Integrated Pipe Corridors

Abstract

Abstract Water leakage is a major problem in underground integrated corridors. Research on damage detection in pipelines using automatic techniques has yielded considerably high detection accuracy. Many proposals in this domain show the presence of leakage only when a certain degree of damage occurs. However, in practical scenarios, the corridor managers of a governing body need to take effective action as soon as a leak is detected. This study makes three contributions to relevant research. First, a state-of-the-art intelligent leakage detection method is proposed for underground integrated corridors using infrared thermal imaging. Considering the difference in terms of specific heat capacity between water and the material of the pipe, when slightly heated, cracks in the loose water-pipe interface can be identified in two-dimensional infrared images. The crack-causing leakage can be extracted via the temperature difference between the leak and the ambient field through image processing based on the grayscale threshold segmentation technique and fractal feature theory. Finally, experiments show that the proposed method can achieve an accuracy of up to 90 %, and compensates well for deficiencies in the traditional method. It can implement intelligent leakage detection in underground integrated corridors at low cost with high efficiency.