猴山岳步道邊坡的滑動方向分析

Sliding Direction Analysis at a Landslide Site near the Mt. Houshanyue Hiking Trail

洪政豪(Cheng-Hao Hung)

指導教授 ： 陳偉堯

國立臺北科技大學/工程學院/土木與防災研究所/碩士(2012年)

摘要

台灣地區每遇颱風及梅雨季節往往造成許多坡地災害發生，導致土石流以及邊坡滑動產生，人民生命財產安全損失甚巨。而產生崩塌滑動地區，往往因為不知道邊坡滑動的確切方向，影響進一步的分析和研究。
本研究針對台北市文山區猴山岳步道台電345kv-039塔柱旁之邊坡，因為在2008年至2009年間，卡玫基、鳳凰及辛樂克等颱風所帶來的豪大雨量造成猴山岳研究區域產生大規模的邊坡滑動。因此，針對邊坡滑動區域使用民國95年、民國97年和民國98年等不同時期之航照影像進行影像分析，利用眾人分工智慧的方式，從98年崩塌後拍攝的航照影像對邊坡滑動區域及滑動區域外圍之樹木挑選出44棵明顯且可辨識的樹木進行樹木輪廓數化，提供給30位參與的測驗人員對照崩塌前95年和97年的航照影像，找出樹木在滑動前的原始位置，以樹木的移動距離以及移動方向作為研究區域的邊坡滑動方向之依據。分析的結果顯示，樹木的移動方向多偏向於北北東，滑動的最大距離約為20至25公尺。應用質點影像分析(Particle Image Velocimetry, PIV)技術，從航照影像上比較崩塌前後時期的水平位移量的變化，最後比較使用PIV分析結果是否與眾人分工智慧分析的結果相似，並歸納兩種分析方式之優缺點。

關鍵字

邊坡 ； 滑動方向分析 ； 眾人分工智慧 ； 質點影像分析

並列摘要

There used to be many disasters happing in the reason of typhoon and plum rain in Taiwan area, which caused many slope disasters such as debris flow and slope sliding, leading to huge loss and threat of safety for people. Since the sliding direction of the slope used to be unknown, accurate analysis and research is thus unavailable.
This study took the slope beside the 345kv-039 tower column of Taiwan Power Company alongside the trail of Mt. Houshanyue in Wenshan District in Taipei. A large-scale slope sliding happened in the subject area, Mt. Houshanyue, due to enormous rainfall brought along by typhoons Kalmaegi, Phoenix, and Sinlaku during 2008 and 2009. Therefore, this study used aerial pictures of the slope sliding area taken during 2006, 2008, and 2009 to conduct image analysis in the way of Mechanical Turk.
Here, 44 obvious and recognizable trees in the aerial pictures taken after the slope sliding were selected for contour digitalization which were then presented to 30 participants for contrasting with the same trees in the aerial pictures taken in 2006 and 2008, in order to find their original location before sliding. Subsequently, sliding displacement and direction of the tress were employed as foundation for research on slope sliding direction of the subject area. The result showed that these trees tended to move towards NNE with maximum displacement around 20 to 25m. The technology of Particle Image Velocimetry (PIV) was applied to figure out displacement variation before and after slope sliding by contrasting the aerial pictures. Finally this study compared the result of PIV analysis with that generated by Mechanical Turk, and summed up advantage and insufficiency.

並列關鍵字

Slope ； Analysis of sliding direction ； Mechanical Turk ； Particle Image Velocimetry

參考文獻

[17] 張詠詠 (2009)，潛移深層地滑邊坡案例之變形特徵與運動歷程探討，碩士論文，國立台北科技大學資源工程研究所。

[22] 趙宥睿(2012)，數值地形模型精度評估及應用-以猴山岳地滑區為例，碩士論文，國立臺北科技大學。

[16] 呂真謀，李明靜，賴泉基，詹勳全，林國暉 (2008)，影像分析法應用於土石流表面速度計算之研究，農業工程學報，第54卷，第1期。

[18] 曾木崧 (2011)，以空照應用PIV探討大安溪蘭勢溪大橋河段變遷，碩士論文，國立交通大學土木工程學系。

[23] 蔡維哲 (2004)，應用航空影像分析於邊坡位移場之方法研究，碩士論文，國立台灣大學土木工程研究所。