Full metadata record
DC FieldValueLanguage
dc.contributor.author賴羿齊zh_TW
dc.contributor.author張憲國zh_TW
dc.contributor.authorLai ,Yi-Chien_US
dc.contributor.authorChang, Hsien-Kuoen_US
dc.date.accessioned2018-01-24T07:39:16Z-
dc.date.available2018-01-24T07:39:16Z-
dc.date.issued2017en_US
dc.identifier.urihttp://etd.lib.nctu.edu.tw/cdrfb3/record/nctu/#GT070351250en_US
dc.identifier.urihttp://hdl.handle.net/11536/140431-
dc.description.abstract本文提出擷取衛星影像的水線並平移至灘線的方法,並以外傘頂 洲為例,探討其灘線變遷。影像處理的技巧包括提高多頻譜態影像解 析度的 IHS 轉換法,區分陸域及海域界線的 NDWI,能提高海陸區域 對比度的影像強化,消除水面波紋及碎波白泡沫的形態學,擷取海陸 交接水線的 Canny 法。因外傘頂洲無潮汐觀測資料,本文以潮位資料 融合法配合 NAO.99b 的潮位數值模式,獲得外傘頂洲精準的潮位。 透過潮位及實際底床灘面的坡度,本文方法可計算出影像的水線平移 至灘線距離,而可決定灘線位置。在選擇斷面上平移後灘線與實測灘 線的 RMSE 為 15.40~53.48m 及 Bias 為 0.65~10.2m,都比原本水線的 值分別修正 44%~74%,及 89%~98%,證實本方法的可行性。 應用本文方法探討外傘頂洲自1994年至2015年間灘線及陸上面 積的變化特性。由分析結果發現外傘頂洲以北端為支點南端逆時針旋 轉,未來會陸化與台灣接連,陸上面積長期有縮小的趨勢,本文預估 外傘頂洲將於 2062 至 2066 間年全部沒入平均海平面以下。然而,外 傘頂洲的陸地面積在夏季是增加的而冬季減少,有季節性的變化特 性,此滿足 Sunamura and Horikawa(1974) 判斷海灘剖面的分類。zh_TW
dc.description.abstractImage processing of detecting waterline in satellite images and waterline-to-shoreline shifting method were proposed in this thesis and its application to investigating shoreline evolution and variation of land area of Waisanding barrier. The image processing includes (1) IHS conversion method for improving the resolution of multi-spectral image (2) NDWI index for determining possible interface of land and sea zones (3) Image enhancement method for improving the contrast between land and sea (4) Morphology for eliminating optional spots such as white ripples of breakers (5) Canny edge detection for capturing the waterline. Due to no observation on tides at Waisanding barrier data blending method associated with NAO.99b tidal model is used to compute the accurate tidal levels. By the trigonometry the obtained tidal level and beachface slope lead to a horizontal distance for shifting the waterline to the shoreline, of which the position is thus determined. RMSEs, Root Mean Square Error, between the shifted shoreline and the measured shoreline at different profiles are evaluated to be 15.40 - 53.48 m and the corresponding Biases are 0.65 - 10.2 m. The values are much iii smaller than those between the waterline and the measured shoreline at the same profiles. Modification on RMSE and Bias using the determined shoreline improves by 44% - 74% and 89% - 98%, respectively, rather than using the detected waterline. The proposed method is proven to be applicable. The shoreline evolution and variations of land area above mean sea level of Waisanding barrier from 1994 to 2015 are investigated using the proposed method. The result shows that southern Waisanding barrier counterclockwise rotates about the southern fixed end. In the near future Waisanding barrier will move and connect to western Taiwan. Meanwhile the dry land of Waisanding barrier has a long-term decreasing trend and possibly disappear during 2062-2066 depending on the linearly fitted rate. The beach of Waisanding barrier suffers from erosion in the winter and accretes in the summer to have a regular and seasonal variations. The result can be explained by the formula for beach types of Sunamura and Horikawa(1974).en_US
dc.language.isozh_TWen_US
dc.subject外傘頂洲zh_TW
dc.subject灘線變遷zh_TW
dc.subject影像處理zh_TW
dc.subject水線辨識zh_TW
dc.subject邊緣偵測zh_TW
dc.subjectWaisanding barrieren_US
dc.subjectShoreline evolutionen_US
dc.subjectImage processingen_US
dc.subjectDetection of waterlineen_US
dc.subjectEdge detectionen_US
dc.title衛星影像的水線辨識及其應用在外傘頂洲的灘線變遷zh_TW
dc.titleWaterline detection from satellite images and its application to the shoreline evolution of Waisanding barrieren_US
dc.typeThesisen_US
dc.contributor.department土木工程系所zh_TW
Appears in Collections:Thesis