A Study of Construction Diagnose - The Appliction of image processing to The Wall Crack Diagnose

• Main Authors: CHANG HUNG-HSIANG, 張宏祥
• Other Authors: TSERNG HUI-PING
• Format: Others
• Language: zh-TW
• Published: 1998
• Online Access: http://ndltd.ncl.edu.tw/handle/27177821374640516995

碩士 === 國立臺灣大學 === 土木工程學研究所 === 87 === Owing to the limited land source and a systematic model for management is lacking in the relevant research on building''s life-cycle and the building forensic problems, construction diagnosis is expected to be one of the leading trends with regard to the construction management. Given the feature of all construction problems, it can generally develop three types of building functions: security function, endurance function, and usage function. Construction diagnosis can be divided into several steps: (1) preliminary diagnosis using the check list or post-occupancy evaluation as a means to understand a building''s general situation; (2) detailed diagnosis making a further diagnosis aimed at building''s function according to the preliminary diagnosis, including problem phenomena, diagnosis methods, and remedy methods; (3) remedies taking the factors such as economy, time, and operability into account while making a strategy based on the detaile Identification of the crack type is crucial in the process of crack diagnosis. This study is making use of the technology of image disposition, Turbo C Program Language and neural network theory to establish the system of crack identification. By the result, we can obtain other information as to the crack''s length, width, and density by which is helpful in the technical aspect of crack diagnosis as well as in making engineering price evaluation.