以作者查詢圖書館館藏 、以作者查詢臺灣博碩士 、以作者查詢全國書目 、勘誤回報 、線上人數:30 、訪客IP:18.188.91.223
姓名 莊雲翰(Yun-Han Chuang) 查詢紙本館藏 畢業系所 土木工程學系 論文名稱 結合影像區塊及知識庫分類之研究-以IKONOS衛星影像為例
(A research on combining Image Segmentation and Knowledge-based Classification using IKONOS images)相關論文 檔案 [Endnote RIS 格式] [Bibtex 格式] [相關文章] [文章引用] [完整記錄] [館藏目錄] [檢視] [下載]
- 本電子論文使用權限為同意立即開放。
- 已達開放權限電子全文僅授權使用者為學術研究之目的,進行個人非營利性質之檢索、閱讀、列印。
- 請遵守中華民國著作權法之相關規定,切勿任意重製、散佈、改作、轉貼、播送,以免觸法。
摘要(中) 相較於傳統的影像判識技術著重於對像元光譜值的分析,在本研究的知識庫系統則是採用物件導向的概念來設計,較接近人的辨識過程且具有遺傳性的特性,適合用於分析類別較複雜的影像。本研究的整體處理過程如可概略的分成影像分割與知識庫判釋兩部分;在影像分割的過程中可同時加入光譜與形狀的因子加以約制,原始影像的像元經過影像分割的步驟之後形成許多區塊所組成的區塊影像,每一區塊具有其光譜以及形狀上的屬性。產生區塊影像之後,再利用知識庫來進行辨識的動作,知識庫的設計亦可加入形狀、光譜、遺傳性等條件來幫助判釋,針對影像上的每個區塊計算其在每一類別之歸屬值,完成判釋的工作。本研究是採用IKONOS衛星影像為測試資料,以同一地區的航照人工判識地物分類作為檢核的依據,成果顯示約有80%的分類正確率。 摘要(英) The traditional methods for image interpretation and analysis are based on the spectral characteristics of single pixel. A general idea of object -oriented programming is used in the method we proposed here. Compare to the traditional methods, the object-oriented processing of image information is nearer to the human cognitive process. Furthermore, the hierarchy between objects is well suit for analysis of complex scenes. The image interpretation process consists of two steps, “image segmentation” and “knowledge-based interpretation”. The segmentation will group pixels of image into numerous regions, which form the object image. We can bind the regions of the image by using combination of spectral and spatial factors in segment process. Each region has unique spectral and spatial characteristics. After segmentation, further classification can be achieved by calculating shape, spectral and hierarchy memberships of each region that belongs to each class in the knowledgebase. The test data of this study is an IKONOS satellite image. Using manual-interpreted airplane photo at the same area as reference for accuracy assessment. The correct classified accuracy of this study is about 80 percent. 關鍵字(中) ★ 知識庫
★ 影像分割
★ 物件導向關鍵字(英) ★ knowledgebase
★ image segmentation
★ object-oriented論文目次 目錄........ I
圖目錄......III
表目錄...... VI
中文摘要.....VII
Abstract....VIII
第一章序論.....1
1-1 研究動機與研究目的...1
1-2 文獻回顧..............2
1-3 章節簡介.............10
第二章研究方法.............11
2-1 影像處理流程與架構.....11
2-2 影像分割...............12
2-3 知識庫系統.............20
第三章測試影像資料與成果...28
3-1 IKONOS 衛星簡介........28
3-2 測試資料...............30
第四章分類成果精度分析.....58
4-1 土地利用真值圖.........58
4-2 其他分類方式之分類成...63
4-3 分類成果之精度分析.....66
4-4 分類成果討論...........73
第五章結論與建議...........75
參考文獻...................76
附錄A 影像分割操作說明.....78
附錄B 知識庫操作說明.......80
附錄C 土地利用分類圖類別合併表......84參考文獻 中央大學土木工程研究所,中壢
羅秋月,2002,”IKONOS 衛星影像正射改正之研究”,碩士論文,國立
中央大學土木工程研究所,中壢
DEFINIENS AG , 2000. eCoginition user’s guide.
http://www.definiens.com ( July,2001)
Hofmann, P., “Detecting urban features from IKONOS data using an
object-oriented approach”, In: Remote Sensing & Photogrammetry
Society (Editor): Proceedings of the First Annual Conference of the
Remote Sensing & Photogrammetry Society 12-14, pp.79-91, September
Hofmann, P., “Detecting Informal Settlements from IKONOS image data
using Methods of object oriented image analysis – an example from
Capetown (SOUTH AFRICA)”, JÜRGENS, Remote Sensing of Urban Areas,
pp.107-117, 2001.
Morisette, J. T. and Khorram, S., “ Accuracy Assessment Curves for
Satellite-Based change Detection”, Photogrammetric Engineering &
Remote Sensing, Vol. 66, No. 7, pp.875-880, July 2000.
Pesaresi, M., and Benediktsson, J. A., ”A New Approach for the
Morphological Segmentation of High-Resolution Satellite Imagery”,
IEEE translations on Geoscience and Remote Sensing,
Vol. 39, No. 2, pp. 309-319, Feb. 2001
Richards, J. A., ”Remote Sensing Digital Image Analysis”,
Springer-Verlag Berlin Heidelberg, 1986.
Schowengerdt, R. A., “Remote Sensing models and methods for image
processing”, 2nd , Academic Press, 1997.
Tonjes, R., Growe, S., Bucker, and Liedtke, C.-E., ”Knowledge-Based
Interpretation of Remote Sensing Images Using Semantic Nets ”,
Photogrammetric Engineering & Remote Sensing, Vol. 65, No. 7,
pp.811-821, July 1999.
Takashi, K., Hiroichi, E., and Yoshiyuki, K., ”Classification of the SPOT
Image Using Spectral and Spatial Features of Primitive Regions That
Have Nearly Uniform Color”, IEEE translations on Geoscience and
Remote Sensing, Vol.28, No. 4, pp.749-752, July 1990.指導教授 陳繼藩(Chin-Fan Chen) 審核日期 2002-7-13 推文 facebook plurk twitter funp google live udn HD myshare reddit netvibes friend youpush delicious baidu 網路書籤 Google bookmarks del.icio.us hemidemi myshare