Combining the 3d Model Generated by Point Clouds and Thermal Infrared Images to Identify the Detects of the Building Facades

• Main Authors: HONH, JIA-JIAN, 洪嘉鍵
• Other Authors: HUANG, YI-SHOU
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/458n47

碩士 === 朝陽科技大學 === 營建工程系 === 107 === Monitoring the health conditions of buildings is a crucial step to extend the life cycles of buildings, especial for those aged buildings. Those aged buildings always contain lots of defects which will cause permanent damages to buildings unless those weaknesses can be identified and fixed. Recently, there are many ways to efficiently collect the 3D spatial information of an object. To restore a 3D model from the collected point clouds in a short time is feasible and reliable. Thermography is used to record the surface temperature information. The surface temperature information is an important clue to indicate the defects presented in the building because those defects could be found in those areas whose surface temperatures are different with their surrounding neighbors. The proposal offers a schematic approach to integrate point clouds collected by various sensors and the segmented thermography to establish a 3D visual system for examining the defects presented in building facades. Point clouds from different sensors offer the various details of the scanned objects. Adopting Unmanned Aerial Vehicle to measure and match point clouds (by Pix4D), focusing on the object to conduct surround shot to establish a more integrated 3d model. Using thermography to conduct nondestructive testing often disturbed by environmental factors such as shadow and reflection; the proposal introduce multiple constant multiplication to form results by Level set numerical solution to enhance to validity of nondestructive testing on thermography; Defects are usually presented in those areas whose surface temperature are different with their neighbors, by this, the potential defects can be identified. The segmented results are projected to the 3D model generated by registering different point clouds. The proposed system offers a 3D model with the detected defects such that different repair methods can be applied to those aged buildings such that the life cycles of buildings can be extended.