關鍵字：影像式建模、相機率定、焦距
HTC butterfly
5. 結論與建議
Douskos, V., Grammatikopoulos, L., Kalisperakis, I., Karras, G. and Petsa, E., 2009. FAUCCAL: an open source toolbox for fully automatic camera calibration. XXII CIPA Symposium on Digital Documentation, Interpretation & Presentation of Cultural Heritage.
Furukawa, Y., Curless, B., Seitz, S.M., and Szeliski, R., 2010. Towards internet-scale multi-view stereo. Computer Vision and Pattern Recognition (CVPR): 1434-1441.
Furukawa, Y. and Ponce, J., 2010. Accurate, dense, and robust multi-view stereopsis. Pattern Analysis and Machine Intelligence, 32(8):
1362-1376.
Lee, P.H., Huang, J.W., and Lin, H.Y., 2012. 3D model reconstruction based on multiple view image capture. Intelligent Signal Processing and Communications Systems (ISPACS): 58-63.
Lowe, D.G., 2004. Distinctive image features from scale-invariant keypoints. International Journal of Computer Vision, 60(2): 91-110.
Rahman, M.H., Qiang, L., Pickering, M., Frater, M., Kerr, D., Bouchey, C., and Delp, E., 2012. Food volume estimation in a mobile phone based dietary assessment system. Signal Image Technology and Internet Based Systems (SITIS):
988-995.
Snavely, N., 2010. Bundler: structure from motion (SFM) for unordered image collections.
http://phototour.cs.washington.edu/bundler/.
Wu, C., 2011. VisualSFM: a visual structure from motion system. http://ccwu.me/vsfm/.
Yang, M.D., Chao, C.F., Lu, L.Y., Huang, K.S., and Chen, Y.P., 2013. Image-based 3D scene reconstruction and exploration in augmented reality. Automation in Construction, 33: 48-60.
Zhang, Z., 1999. Flexible camera calibration by viewing a plane from unknown orientations.
Computer Vision, 1: 666-673.
Zhou, K., 2010. Structure & motion, structure in pattern recognition. Vienna University of Technology, faculty of informatics, Institute of
Computer Graphics and Algorithms, Pattern Recognition and Image Processing Group.
1Master, Department of Civil Engineering, National Chung Hsing University Received Date: May 25, 2015
2 Professor, Department of Civil Engineering, National Chung Hsing University Revised Date: Jul. 27, 2016
3 Ph.D., Student, Department of Civil Engineering, National Chung Hsing University Accepted Date: Sep. 01, 2016
* Corresponding Author, E- Mail: [email protected]
Image-based modeling for multi-source images and establishing an accuracy index of reconstruction
Wei-Ting Pan1 Ming-Der Yang2* Chih-Fan Chao3
Abstract
Most images for Image-Based Modeling (IBM) were taken from high resolution cameras in the past. With its widespread, cameras on smart mobile devices can be one type of image sources. The aim of this study is to build up an index of reconstruction quality which can calculate the error of the volume through the focal length.
Thus, the index can estimate the accuracy of reconstructed models rapidly by multi-source images. This study uses various kinds of smartphones and a digital single lens reflex (DSLR) to take photographs of concrete specimens and to reconstruct 3D models. Focal lengths calculated by camera calibration are considered as the true value to estimate focal lengths accuracy by structure from motion (SfM) and then the volume error of the 3D models. Finally, the accuracy index of reconstructed models was established through a linear regression based on focal lengths and volume error. Owing to unknown true volume in most actual conditions, it becomes convenient to apply various mobile devices with the camera to take photographs and reconstructed models.
According to the accuracy index, the error of earthwork quantity (Y) can be estimated by focal length (X), and the result can be the basis of disaster analysis and prevention measures aprior.
Keywords: Image-Based modeling, camera calibration, linear regression
本學刊定期出版攝影測量及遙感探測領域之學術論著、技術短文及應用性報導。所收錄之稿件 皆須通過本編輯委員會安排之專家學者審查程序，並由通訊作者簽署著作權讓渡書。作者可自行聲 明文稿屬學術論著、技術短文或應用性報導之類別，但本編輯委員會將依據審查委員意見決定收錄 文稿之類別。本學刊不接受已經刊載過之文稿，於研討會發表過之論文，經適當增補其內容者為例 外。
投稿之電子郵件信箱地址為 [email protected]，請以電子郵件附件方式寄送稿件之 WORD 檔及 PDF 檔各一份。為便利編審作業，請於稿件 PDF 檔中移除作者之相關資料（如姓名、服務機關、
電話、及電子郵件信箱等）。稿件撰寫要點如下：
1. 頁面：採用 A4 規格頁面。
2. 字體：(1) 中文均以新細明體橫式打字；(2) 英文均以 Times New Roman 字體橫式打字。
3. 單位：內文、圖表及符號中之度量單位皆採用 SI 公制單位。
4. 首頁：首頁應含內容依次為「題目、作者姓名、摘要、關鍵詞、及作者服務機關及職稱（置於 頁尾）」，接續文稿內文。
5. 末頁：末頁應含內容與首頁相同，但皆以英文書寫。
6. 題目：論文題目宜簡明，字體大小為 20，英文題目每字第一位字母以大寫字體打印。
7. 作者：作者姓名列於題目之下方，字體大小為 14，其服務機關及職稱以1, 2,…上標編號註記於頁 尾，字體大小為 12。
8. 摘要：摘要字數以 250 字為限，字體大小為 12。
9. 關鍵詞：列出三至五個關鍵詞，字體大小為 12。
10. 章節及標題：(1)章節標題置中，編號一律為阿拉伯數字，如 1，2，...，標題字體大小為 18；(2) 小節標題靠左，編號一律為 1.1，1.2，1.1.1，1.1.2……，字體大小為 14。
11. 內文：內文之長度(含圖表)以不超過 25 頁為原則，各段落第一行應縮排，字體大小為 12。
12. 數學式：所有數學式均須標式號於圓括弧內，數學式內所有符號須於符號第一次出現時，應於 數學式下方逐一加以定義。
13. 圖及影像：(1)圖及影像均須數位化置於文稿中適當位置；(2)所有圖及影像下方須附圖流水編號 及說明；(3)本刊接受彩色圖及影像，但紙版印刷一律採灰階印刷。
14. 表格：(1)表格均須數位化置於文稿中適當位置；(2)所有表格上方須附表流水編號及說明；(3) 本刊接受彩色表格，但紙版印刷一律採灰階印刷。
15. 參考文獻：(1)參考文獻應列於文末，文稿中未直接引用者不得列入；(2)先排列中文參考文獻，
再排列英文參考文獻，其他文字參考文獻列於英文參考文獻之後；(3)中文參考文獻以第一作者 姓名筆劃順序排列，英文參考文獻以第一作者之姓依英文字母順序排列。當作者相同時，依發 表年代排列，作者及發表年代完全相同時，依日期在年代後加以 a, b, c 順序。第一作者姓名相 同時，先排列單一作者之參考文獻，其次為多作者文獻；(4)參考文獻之格式如下列：
期刊論文：作者姓名，年代。文章標題，期刊名稱，卷數(期數)：起始頁數─結束頁數。
書籍：作者姓名，年代。書名，出版者，出版地。
論文集：作者姓名，年代。篇名，論文集名稱，發表地點，卷(期)：起始頁數─結束頁數。
16. 參考文獻引用：(1)作者為一人時，以作者全名(中文文獻)或姓(英文文獻)及發表年代，如李良輝 (1991)或(李良輝，1991)、Jones (1979)或(Jones, 1979)；(2)作者為二人時，以二作者全名(中文文 獻)或姓(英文文獻)及發表年代，如施學延及陳良健(2011) 或(施學延及陳良健，2011)、Jones &
Smith (1979)或(Jones & Smith, 1979)；(3)作者為多人時，以第一作者全名(中文文獻)或姓(英文文 獻)，之後加「等」(中文文獻)或「et al.」(英文文獻)及發表年代，如(郭寶章等，1976)或郭寶章 等(1976)、Jones et al. (1979)或(Jones et al., 1979)；(4)作者及發表年代完全相同時，依日期在年 代後加以 a, b, c 引用，如 Jones & Smith (1979a)、Jones & Smith (1979b)。
The Journal of Photogrammetry and Remote Sensing periodically publishes academic articles, technical short papers and application reports addressing topics in photogrammetry and remote sensing. Acceptance of submitted manuscripts for publication is based on a peer-reviewing process arranged by the editorial board. A manuscript will not be formally accepted until a copyright agreement is received. Although authors can indicate their submitted manuscripts belonging to one of the categories of academic articles, technical short papers or application reports, the editorial board will make the final decision based on the suggestions of reviewers. This journal does not accept any paper has ever been published, but a paper only published in conference proceedings and revised appropriately is an exception.
The E-mail address for manuscript submission is [email protected]. Please send your manuscript in both WORD and PDF formats via e-mail attachments. In the WORD file, please provide author information. However, for blind reviewing, please eliminate author information in the PDF file.
Instructions for authors to prepare a manuscript are as follows:
1. Typing format: Manuscripts should be typed single-column and double-space on A4 page size using font-type of 12pt Times New Roman. The pages should be numbered.
2. Units: Manuscripts must use SI (Metric) units in text, figures, or tables.
3. Cover page: The cover page should contain manuscript title, author names, abstract, key words, and author affiliations.
4. End page: The end page contains the same information as the cover page, but in Chinese.
5. Title: The title of the manuscripts should be concise and informative. Please use font size of 20pt and capitalize the initial letter of each word in the title.
6. Authors: Authors names (using font size of 14pt) are listed below the title, affiliations (using font size of 12pt) are listed at the end of the first page marked with 1, 2,....
7. Abstract: The abstract should be shorter than 250 words and typed using font size of 12pt..
8. Keywords: Keywords (no more than 5 words) should be given and placed below the abstract.
9. Sections and Sub-sections: Section heading (using font size of 18pt) should be centered and numbered in 1, 2… etc. Sub-sections (using font size of 14pt) are justified to the left in 1.1, 1.2, 1.1.1, 1.1.2…etc.
10. Paper length: Published papers are limited to 25 pages.
11. Equations: Type formulae and equations clearly in black and numbered in curved brackets. All symbols in equations should be clearly defined when they firstly appear in the manuscripts.
12. Figures and Tables: Figures and tables are numbered and described in short. When papers are accepted, original figures, tables, and photographs are expected for reproduction. Color figures and tables are acceptable in digital version, but they will be printed in gray levels in hard copies.
13. References: Only works cited in the text will be included. References should be cited in the text by giving the last name of the author(s) and the year of publication of the reference, such as Lee and Huang (1979) or (Lee and Huang, 1979). With three or more names use the form “Huang et al. (1995) or (Huang et al., 1995)” in the text. When two or more references have the same text identification, please distinguish them by appending “a”, “b”, etc., to the year of publication. References should include the complete title(s), the number of the volume and the page(s) of the information cited. References should be listed in alphabetical order, according to the last name of the first author, at the end of the paper.