現在位置首頁 > 博碩士論文 > 詳目
論文中文名稱:應用紋理輔助分析物件導向式分類於作物判釋之研究 [以論文名稱查詢館藏系統]
論文英文名稱:Application of Texture-Assisted Analysis of Object-Oriented Classification to Crop Interpretation [以論文名稱查詢館藏系統]
院校名稱:臺北科技大學
學院名稱:工程學院
系所名稱:土木工程系土木與防災碩士班
畢業學年度:106
畢業學期:第二學期
出版年度:107
中文姓名:馬儷瑞
英文姓名:Lee-Ray Ma
研究生學號:105428041
學位類別:碩士
語文別:中文
口試日期:2018/07/30
論文頁數:86
指導教授中文名:朱子偉;譚智宏
指導教授英文名:Tzyy-Woei Chu;Chih-Hung Tan
口試委員中文名:陳莉;阮忠信;朱子偉;譚智宏
中文關鍵詞:灰階共生矩陣物件導向式分類WorldView-2作物判釋紋理分析
英文關鍵詞:Gray Level Co-Occuruence MatrixObject-Oriented ClassificationWorldView-2Crop InterpretationTexture Analysis
論文中文摘要:台灣地區地狹人稠,導致土地利用相對複雜,因此大面積的地面資料獲取不易,而過去以人力調查方式,不僅耗力費時,亦十分不經濟。隨著遙感探測的進步,利用影像自動化分類來獲得土地覆蓋的資料已漸成趨勢。由於遙測影像空間解析度的提升,影像提供的紋理(texture)資訊也相當豐富,而傳統的以單一像元(pixel)為分類單元的像元式分類,已無法精準表達複雜的土地利用特性,所以逐漸發展為以多邊形為分類單元的物件導向式分類(Object-Oriented Classification)。故本研究旨在分析高解析度衛星影像,運用物件導向式分類結合紋理資訊於複雜的農地進行作物判釋,以提升準確度並作為後續相關研究及政策的參考依據。
本研究應用紋理分析法中的灰階共生矩陣(Gray Level Co-Occurrence Matrix, GLCM),並結合光譜空間中的影像角二階矩、對比度、相關性、同質性、不相似性、熵及平均之紋理特性參數計算,作為建立不同作物類別之紋理資料之參考。研究採用高解析度衛星WorldView-2攝於桃園市平鎮區中的石門農田水利會所屬山溪第一輪區的影像資料,分別以像元式分類、物件導向式分類及物件導向式分類結和紋理資訊三種方法進行座誤判式,並比較其準確度。
研究結果顯示,像元式分類、物導向式分類及物件導向式分類結和紋理資訊,其整體精度依序分別為 78.2%、84.96%和93.23%,而Kappa值分別為0.7233、0.8078、0.914。結果顯示,加入紋理資訊後,的確可以提高分類準確度。然而由於各個作物類別內的紋理性質具有差異性,在判釋的時候紋理之間會相互影響,若加入超過一個以上的紋理資訊可能會導致干擾而降低了分類的精度,造成反效果。
論文英文摘要:Mixed land use as well as densely populated areas in Taiwan generally results in the relatively complicated land cover information. Therefore, it seems arduous to acquire the extensive spatial ground information through field survey given that not only is it time-consuming but it is also expensive. As remote sensing technology advances, the growth in utilization of automatic image classification to obtain land cover data becomes obviously prevailing. In addition, the texture information within the image has become rich and enhanced by the aids of the elevated spatial resolution of the telemetry image. On the contrary, the traditional pixel based classification using a single pixel as the classification unit is unable to accurately present the complex land use characteristics. Accordingly, the technology has developed into an object-oriented classification with polygons as the unit. Thus, this study aims to employ an approach using object-oriented classification combined with texture information in analyzing high-resolution satellite imagery on complex agricultural land for crop interpretation. Its objective is to improve the accuracy of image classification on crop interpretation.
This study applies the Gray Level Co-Occurrence Matrix (GLCM) method in the texture analysis, and combines seven texture parameters in the spectral space (Ang Second Moment, Contrast, Correlation, Homogeneity, Dissimilarity, Entropy, and Mean) to perform calculations and provide texture information for different crops as further references. Additionally, the high-resolution image of Shansi first rotation block in Taoyuan City, taken by WorldView-2 satellite, was conducted in this study. Three approaches of image classification including pixel based, object-oriented, and object-oriented combined texture information are accomplished and evaluated their precision in crop interpretation.
The results show that the overall accuracies and Kappa values of pixel classification, object-oriented classification and object-oriented classification with texture information are 78.2 % and 0.7233, 84.96 % and 0.8078, and 93.23 % and 0.914, respectively. It is concluded that the addition of texture information in object- oriented classification achieves the best performance and improves a lot in overall accuracy than traditional pixel-based approach. However, every type of crop matches only one specific texture parameter to obtain the best accuracy. It should be cautiously aware that utilization of more than one texture information does induce mutual interference and reduce the overall accuracy.
論文目次:摘 要 i
ABSTRACT iii
誌 謝 v
目 錄 vi
表目錄 viii
圖目錄 ix
第一章 緒論 1
1.1研究背景 1
1.2研究動機及目的 2
1.3研究流程 2
第二章 文獻回顧 5
2.1 遙感探測 5
2.1.1 成像原理 5
2.1.2 灰階影像 6
2.2 遙感探測相關文獻 7
2.3 物件導向式分類相關文獻 8
2.4 紋理分析相關文獻 10
第三章 理論與方法 16
3.1 研究區域概述 16
3.2研究資料特性與收集 17
3.2.1常見衛星特性介紹 17
3.2.2 WorldView-2衛星介紹 20
3.2.3 地真資料 22
3.3物件導向式分類 23
3.4分類門檻值 31
3.4.1紋理 32
3.4.2 灰階共生矩陣 32
3.6 評估方法 38
第四章 結果與討論 41
4.1研究類別 41
4.2 光譜反應值分類 43
4.3 紋理資訊輔助分析結合光譜值分類 51
4.3.1 紋理參數的選定 51
4.3.2紋理參數值的分佈 63
4.3.3小結 71
第五章 結論與建議 75
5.1 結論 75
5.2 建議 76
參考文獻 78
附錄:檢核點坐標 83
論文參考文獻:1. Douglas A. Stow, Dong Mei Chen, “Sensitivity of Multitemporal NOAA/AVHRR Data of an Urbanizing Region to Land-Use/Land-Cover Changes and Misregistration,” Remote Sensing of Environment, vol. 80, no. 2, 2002, pp. 297-307.
2. L. Ravanel, F. Magnina, P. Deline, “Impacts of the 2003 and 2015 Summer Heatwaves on Permafrost-Affected Rock-Walls in the Mont Blanc Massif,” Science of The Total Environment, vol.609, no.31, 2017, pp.132-143.
3. Kelly M, D. Shaari, Q. H. Guo, and D. S. Liu, “A Comparison Ofstandard and Hybrid Classifier Methods for Mapping Hardwood Mortalityin Areas Affected by Sudden Oak Death,” Photogrammetric Engineeringand Remote Sensing, vol. 70, no. 11, 2004, pp. 1229-1239.
4. Oruc, M., Marangoz, AM, & Buyuksalih, G., “Comparison of pixel-based and object-oriented classification approaches using Landsat-7 ETM spectral bands,” Annual Conference, Istanbul, 2004, pp. 19-23.
5. D. Alfred Hancq, Andrew J Walters and Jack.L Beuth, “Development of an Object-Oriented Fatigue Tool,” Engineering with Computers, vol.16, no. 2, 2000, pp. 131-144.
6. Hassan Fathizad, Mohammad Ali Hakimzadeh Ardakani, Ruhollah Taghizadeh Mehrjardi, Hamid Sodaiezadeh, “Evaluating Desertification Using Remote Sensing Technique and Object-Oriented Classification Algorithm in the Iranian Central Desert,” African Earth Sciences, vol.145, no.9, 2018, pp.115-130.
7. H. Kaizer, A Quantification of Textures on Aerial Photographs, Boston: Univ. Res. Lab, 1955, pp. 121.
8. Gonzalez, R.C., Woods, R.E., Eddins, S.L., Digital Image Processing Using MATLAB, New Jersey: PHI Pearson, 2004, pp.23-35.
9. Jack Sklansky, “Image Segmentation and Feature Extraction,” Transactions on Systems, Man, and Cybernetics, vol. 8, no. 4, 1978, pp. 237-247.
10. Haralick RM, Shanmugan K and Dinstein I, “Textural Features for Image Classi-fication,” Transactions on Systems, Man and Cybernetics, vol. 3, no. 6, 1973, pp.610–621.
11. Trimble Documentation:eCognition Developer 9.0 User guide, 2014.
12. Weszka,J.S., C.R. Dyer, and A. Rosenfeld, “A Comparative Study of TextureMeasures for Terrain Classification,” Trans. on Syst., Man, and Cybern., vol. 6, no.4, 1976, pp.269-285.
13. Conners, R.W., and C.A. Harlow, ”A Theoretical Comparison of Texture Algorithms,” Trans. on Pattern Analysis and Machine Intelligence, vol. 2, no. 3, 1980, pp. 204-222.
14. J.M.H, Du Buf.M, Kardan, M.Spann, “Texture Feature Performance for Image Segmentation,” Pattern Recognition, vol.23, no.3-4, 1990, pp.291-309.
15. Philippe P.Ohanian, Richard C.Dubes, “Performance evaluation for four classes of textural features,” Pattern Recognition, vol. 25, no.8, August 1992,pp. 819-833.
16. Cottin, A. G., D. L. Forbes, and B. F. Long, “Shallow seabed mapping and classification using waveform analysis and bathymetry from Shoals lidar data,” Canadian Journal of Remote, Sensing, vol. 35, no. 5, 2009, pp. 422-434.
17. C. Arcidiacono , S.M.C. Porto, “Pixel-Based Classification of High-Resolution Satellite Images for Crop-Shelter Coverage Recognition,” Acta Horticulturae, vol. 937, no.5, 2010, pp. 1003-1010.
18. Marceau, D. J., P. J. Howarth, J. M. Dugois and D. J. Gratton, “Evaluation Of The Grey-Level Co-Occurrence Matrix Method for Land-Cover Classification Using SPOT,” Transaction on Geosciences and Remote Sensing, Imagery, vol. 28, 1990, pp.513-519.
19. J. Tetuko, R. Tateishi and K. Wikantika, “A Method to Estimate Tree Trunk Diameter And its Application to Discriminate Java-Indonesia Tropical Forests,” International Remote Sensing, vol. 22, no. 1, 2001, pp.177-183.
20. Fu Zheng, Zhang Hua-Shan, Zeng Xiao-Ying and Xiong Hai-Rong, “Characterization and Site-Directed Mutagenesis of Xylanase 1YNA from Thermomyces Lanuginosus DSM10635,” Hubei Agricultural Sciences, vol.50, no. 7, 2011, pp.1487-1493.
21. Wei, Wenxia, Chen, Xiuwan and Ma Ainai, “Object-Oriented Information Extraction and Application in High-Resolution Remote Sensing Image,” Proceeding of the IGARSS 2005 Symposium, Seoul, 2005, pp. 3803-3806.
22. Rouse, J. W., Haas, R. S., Schell, J. A. and Deering, D. W., “Monitoring Vegetation Systems in the Great Plains with ERTS,” Proceedings 3rd ERT S Symposium1, 1973, pp. 48–62.
23. Christopher D, Elvidge and Zhikang Chen, “Comparison of Broad-Band and Narrow-Band Red and Near-Infrared Vegetation Indices,” Remote Sensing of Environment, vol.54, no. 1, 1995, pp. 38-48.
24. R.A Montserud and R Leamans, Comparing Global Vegetation Maps with the Kappa Statistic, Ecological Modelling, vol. 62, no. 2, 1992, pp. 275-293.
25. Jacob Cohen, “A Coefficient of Agreement for Nominal Scales,” Educational and Psychological Measurement, vol.20, no. 1, 1960, pp.37-46.
26. M. Lillesand Thomas, W. Kiefer Ralph and W. Chipman Jonathan, Remote Sensing and Image Interpretation, New York: John Wiley and Sons Publications, 2004, pp.78-80.
27. Fawwaz T. Ulaby, F. Kouyate, B. Brisco and T. H. Lee Williams, “Textural Infornation in SAR Images,” Geoscience and Remote Sensing, vol. 24, no. 2, 1986, pp. 235-245.
28. Ketut Wikantika, A.B. Harto, Ryutaro Tateishi, Wihartini, J. Tetuko S.S. and Jong Hyun PARK, “An Investigation of Textural Characteristics Associated with Spectral Information for Land Use Classification,” Geoscience and Remote Sensing Symposium, Honolulu, 2000, pp. 2915-2917.
29. 陳益凰,應用多時段衛星影像辦識水稻田之研究,碩士論文,國立成功大學測量工程研究所,台南,1998。
30. 蔡光榮、陳建富,「遙測技術應用於台灣中部集集地震災區崩塌地植生處理工程之植生指數評估」,航測及遙測學刊,第十卷,第二期,2005,第203-212頁。
31. 陳誌彬,混合式資訊重建演算法應用於多時期Landsat衛星影像修復,碩士論文,國立成功大學測量工程研究所,台南,2014。
32. 徐百輝大地的辨識密碼高光譜影像。科學發展,第一卷,第四百一十六期,2007,第13-19頁。
33. 陳健雄,具邊緣保留特性之資料隱藏方法-以空間域灰階影像為例,碩士論文,國立成功大學工程科學研究所,台南,2009。
34. 林承毅,空載雷射掃描資料產生數值高程模型之研究,碩士論文,國立高雄應用科技大學土木工程與防災科技研究所,高雄,2004。
35. 莊雲翰,結合影像區塊及知識庫分類之研究—以 IKONOS 衛星影像為例,碩士論文,國立中央大學土木工程研究所,桃園,2002。
36. 林啟陽,「以SuperGIS Toolkit建立空間統計功能之流程設計」,地理資訊系統季刊, 第七卷,第二期,2013,第3-5頁。
37. 戴文彬,自組紋理辦識技術在遙測影像處理之應用,台北:行政院國家科學委員會補助專題研究計畫,2000。
38. 蔡敏之,應用輔助資訊及類神經網路於遙測影像土地利用分類之研究--以識別都市區為例,,碩士論文,國立成功大學測量工程研究所,台南,1997。
39. 黎瑋,紋理分析於遙測影像分類之研究,碩士論文,國立中央大學土木工程研究所,桃園,1997。
40. 陳彥宏,運用紋理資訊輔助高解析度衛星影像於都會區水稻田萃取之研究,碩士論文,逢甲大學土地管理所,台中,2004。
41. 周明中,紋理輔助高解析度衛星影像分析應用於偵測入侵性植物分布之研究,碩士論文,國立中央大學土木工程研究所,桃園,2005。
42. 鄭丁元,高解析度衛星影像於水稻田坵塊萃取之研究,碩士論文,逢甲大學土地管理所,台中,2006。
43. 林榮章,都會區多解像力遙測影像之紋理分析,碩士論文,國立中興大學土木工程研究所,台中,1999。
44. 鄭雅文,高解析度衛星影像分類應用於國土利用調查之研究,碩士論文,國立交通大學土木工程研究所,新竹,2007。
45. 高玉惠,小波轉換應用於影像自動判釋崩塌地分析,碩士論文,國立成功大學測量工程研究所,台南,2004。
46. 廖浩鈞,以QuickBird高解析度遙測影像以分類法判釋南二高邊坡保護工程之研究,碩士論文,國立成功大學測量工程研究所,台南,2006。
47. 唐德誠, 灰度共現矩陣於多波段多極化SAR影像分類之研究,碩士論文,國防大學中正理工學院軍事工程研究所,桃園,2003。
48. 沈佩萱,多個母體的變異係數比較之有母數強韌法,碩士論文,國立中央大學統計研究所,桃園, 2007。
49. 洪凱政,應用多光譜影像多種特徵偵測崩塌地之研究,碩士論文,國立成功大學測量工程研究所,台南,2009。
50. 農業工程研究中心,臺灣石門農田水利會以遙測技術調查灌區內農作物分佈與需水量推估計畫期末報告書,桃園:農業工程研究中心,2016。
51. 灰階影像,維基百科。
論文全文使用權限:不同意授權