Use of Passive Infrared Thermography Method for Subsurface Damage Evaluation of Dam Concrete by Integration of Image Clustering with Heat Balance Analysis

Abstract

In this study, we combine clustering of images obtained by passive infrared thermography method and heat balance analysis for the non-contact evaluation method of subsurface damage of dam concrete. It is possible to classify the concrete, efflorescence, chalk, joint and vegetation from obtained visible and thermal images by image clustering. Heat balance analysis is applied for the classified concrete part, the void and non-damaged surface temperature could be reproduced with an MAE (Mean Absolute Error) of 1.5°C. These results indicate that it is possible to reveal the relationship between concrete surface temperature and subsurface damage situation. This method is expected to apply to the quantitative evaluation of subsurface damage.