基於數位影像識別技術之結構健康監測系統之研究

Abstract

土木基礎建設之使用年限多達數十年或甚至百年以上，在此使用期間之內，結構體 除了承受基本設計的使用年限之外，亦受到周圍環境之影響，如地震、颱風等天然災害 之侵襲，若結構缺乏適當的維修與保養，將造成許多結構物在未達到其使用年限之前， 便已損害而影響其使用性與安全性。因此，吾人亟須建立完善的智慧型防災監測技術， 以提昇基礎建設的安全性與可靠度。現今，智慧感測技術與資訊科技的進展使得結構智 慧化監測變成可能，多數的結構已建置有許多的先進的加速度計、速度計、光纖感測器 等等的感測儀器。然而，在有限經費的考量上，不可能在大範圍佈置大量的感測結點， 且某些構件位置布設感測器也相當的不易，而無線感測網路與數位影像的應用恰可補強 這一部分的機能。本研究團隊在近幾年已針對無線感測網路於結構健康監測進行相關研 究，已有不錯的成果，並已結合電機領域學者開發土木專屬之無線感測平台。而以數位 影像技術為基楚之監測系統，則有價格低廉、精確度高、安裝快速等優點，因此，本研 究將進一步對影像處理技術於土木結構健康監測進行一系列之研究，發展數位影像識別 技術之結構健康監測系統，此系統將可以有效提高土木公共基礎建設之防災監測的能 力。

Civil infrastructures are prone to damage and deterioration during their service life. Damage assessment attempts to determine whether structural damage has occurred, as well as the location and extent of any such damage. Hence, to establish a complete sensor system to detect damage of infrastructures is an essential task. Detecting structural damage and identifying damaged elements in a large complex structure, however, are challenging tasks since the in situ measured data of large civil infrastructures such as bridges and buildings are inaccurate (owing to noise corruption) and often incomplete (for economy consideration). Recently, development of intelligent sensor system for civil infrastructures becomes possible due to the improvement of advanced sensor and ICT technology. For traditional wired sensor system, it is difficult to set a bulky sensor system owing to the limitation of cost and inadequacy of setting. The advantage of wireless senor network and digit image processing technology can overcome the drawback of aforementioned difficulty. Recently, research on developing intelligent wireless sensor network in structural health monitoring (SHM) has been done via cooperating with EE researchers in our research team and a prototype of smart wireless sensor has been developed. Rather than wireless senor system, digit image processing technology has got more attention due to the advantage of low price, high precision, and easy installment. The objective of this research is to study the application of digit image processing based SHM for civil infrastructures. Algorithms for digit image based system identification will be developed. Small scaled building models and real bridge structures will be employed to verify the feasibility of developing digit image based SHM system.