Mobile geo-referenced thermal infra-red imagery recording for leakage detection in urban district heating networks,” Applied Geoinformatics for Society and Environment
Miraliakbari, A., Hahn, M., Engels, J., Sohel, F. and Bennamoun, M. (2013) Mobile geo-referenced thermal infra-red imagery recording for leakage detection in urban district heating networks,” Applied Geoinformatics for Society and Environment. In: Applied Geoinformatics for Society and Environment (AGSE) 2013, 16 - 19 December 2013, CEPT University, Ahmedabad
Localization of heating networks and detection of their leakages are significantly important for energy suppliers. Traditional ways of leakage detection are time consuming and costly. It might sometimes end up with wrong detection. To address such problems, this paper represents a twostep procedure for the leakage detection in thermal images. Localization of the captured data is enabled using a GPS-aided inertial navigation system synchronized with the thermal camera. Leakage and poor insulation detection is achieved in two phases: The first phase is to register the base maps of the network to the thermal sequence and investigate the indicated temperatures surrounded by the network domain (region of interest). The second phase is to cluster the thermal image and look for the cluster of the highest temperature which suggests the heat loss. K-means clustering is selected due to its high speed of computation compared to other clustering methods. The temperatures can be grouped into 4 to 6 clusters. Three major predefined classes correspond to cold, medium and high temperature. The experimental results show that poor insulations can be successfully detected.
|Publication Type:||Conference Paper|
|Item Control Page|
Downloads per month over past year