請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/61345
標題: | 以影像技術進行結構物相對幾何變遷偵測 Image-based technique for monitoring relative deformations of structures |
作者: | Chien-Yi Lai 賴建邑 |
指導教授: | 韓仁毓(Jen-Yu Han) |
關鍵字: | 變形分析,攝影測量,非迭代法,相對方位解算, Deformation Analysis,Photogrammetry,Non-iterative Approach,Relative Orientation, |
出版年 : | 2013 |
學位: | 碩士 |
摘要: | 結構物受外力影響後發生變形,導致使用安全度及年限降低,對人民之生命財產是一大威脅,故如何快速且有效的偵測結構物變形,適時的提出警告並修補以延長其使用壽命為一重要議題。傳統變形監測方法,或多或少都會受限於現地環境或方法特性的關係,導致測量工作受阻,無法進行快速有效率的變形監測工作,攝影測量方法擁有可快速且大量記載物空間資訊的優勢,成為現行常使用的變形偵測方法之一。攝影測量方法的基本原理是透過光線幾何約制條件,找到物像對應關係,而傳統攝影測量解算技術必須於物空間佈設控制點後進行拍攝,接著找到適當的初始值,使用以共線式為基礎之最小二乘法進行迭代求解,因此便利性較低,且若初始值的給予不夠準確,會導致解算成果發散。本研究以解算相機間之相對姿態參數為出發,在解算過程中使用非迭代法進行運算,可在不需要物空間控制點亦不需給予參數初始值的情況下,進行相對姿態參數求取,而後以此相對姿態參數配合已知之相對基線向量,進行空間前方交會後取得物點坐標,以物點間的相對幾何關係作為變形偵測的判識標的,研判變形發生與否,如此一來,便可在滿足精度要求的前提下,將此應用對施作環境的要求降至最低,大幅提升其現地適用性。經數值實驗與實際測試結果得知,本研究所提出之方法可迅速且正確地重構相機之相對幾何關係,因此可作為工程結構物變形監測任務一項便利可靠的分析技術。 Deformation of a structure is one of the most important signs that indicate its safety and usability. How to measure the deformation accurately and efficiently has become a critical issue. Many techniques have been developed to detect the structural deformation in the past few years. Among all methods, photogrammetry technique has its advantage of recording a huge amount of spatial information within a short period of time and is thus a frequent choice in monitoring structure deformations. The main idea of Photogrammetry is to determine the relationship between image space and object space based on optical constraints. Classical approaches require a set of well distributed control points and appropriate initial values before the orientations of images can be determined using an iterative least squares technique. This will limit the efficiency in a real world application. In this study, a non-iterative approach is applied to solve the relative orientations between images. The obtained geometry in the object space is then used for identifying possible deformation signals. Based on the results from numerical tests, it is illustrated that the proposed approach is capable of reconstructing the relative geometry between images in an efficient and accurate manner. Consequently, it can be a practical and reliable solution for structure deformation monitoring tasks. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/61345 |
全文授權: | 有償授權 |
顯示於系所單位: | 土木工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-102-1.pdf 目前未授權公開取用 | 2.45 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。