Integration Of Thermal Imaging And Image Process To Detect Tile Peeling of External Walls

• Main Authors: Yu-Liang Kwei, 歸友亮
• Other Authors: 謝孟勳
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/50460437591394650015

碩士 === 國立中興大學 === 土木工程學系所 === 104 === Tiles for building facades have become mainstream styles of building construction since the Japanese colonial period. According to literature reviews, tile peeling mainly due to poor construction quality or aging adhesive layer. However, after long-term severe variation between high and low temperature, the tiles are apt to peel from the facades and cause serious accidents in Taiwan. Therefore, detecting defective tiles before the tiles fall and hurt/kill pedestrians becomes a very important issue. This study is aim to develop an efficient methodology to detect tiles with back-interspace which are apt to peel soon. The research applied an infrared thermography camera to shoot tile surface. The tiles with solid adhesion or with back-interspace will has a different heat conduction phenomenon, thus they also show significant differences in surface temperatures. An intensive experiment was conducted in building façade of the Department of Civil Engineering, NCHU, where tiles peeled seriously. The infrared thermal images are further process through many image process techniques including K-means algorithm, Contour detection, Gaussian, HSV conversion. The study has successfully developed a Python program integrating with Open-CV module to build an efficient tile peeling detection system.