應用數值航空影像於投89線道崩塌區之位移場分析

Analysis of Displacement Field on Nantou 89th County Road’s Landslides by Using Digital Photogrammetry

黃雅萱(Ya-Hsuan Huang)

指導教授 ： 徐松圻

朝陽科技大學/理工學院/營建工程系/碩士(2016年)

摘要

南投縣仁愛鄉力行產業道路(投89線道)，因其地質狀況脆弱，颱風豪雨來臨時易造成此區域大面積之崩塌。加上地勢陡峭造成監測上之困難，以傳統之監測技術無法完成位移之分析，因此使用質點影像測速技術以求得其位移場。
本研究主要是以質點影像測速技術(PIV)求得崩塌地及其周邊區域之水平位移場。利用南投縣投89線道路之五處崩塌地作為研究對象，將2007、2008、2010、2012年之原始航照，利用ERDAS IMAGINE LPS 模組自行產製DTM及正射影像，並配合PIVview2C軟體進行PIV分析。
在研究區域範圍內，其正射影像之位置準確度主要介於0.50-1.50 m之間。而PIV分析結果大多可合理得到其位移場分布，其位移準確度偏差量介於0.14-4.59m之間，並利用地形變異分析(DoD , Difference of DTM)及地形剖面進行判釋，因PIV是採影像特徵相似性判釋位移向量，因此，大多處分析較差之區域（如32.5K）多是崩坍區受植披、日照陰影影響，分析較為不良，利用地形變異分析、地形剖面可驗證PIV分析結果是否合理，也可補足PIV分析不足。經由PIV分析計算崩塌地之移動量和位移區塊，藉此對投89線道崩塌地的現象與成因提出解釋並歸類整理，期望能對往後的改善工程提供參考。

關鍵字

崩塌地 ； 質點影像測速技術 ； 地形變異分析 ； ERDAS ； 投89線道

並列摘要

Lixing Industrial Road (89th County Road), which located at Ren’ai Township, Nantou, Taiwan. The geological condition within this region is very broken and fragile. It could easily lead to a large area collapse in this region when typhoon rains comes. Due to the steep terrain leads to difficulties of the monitor system, the traditional monitoring techniques can not satisfy the displacement analysis. Thus, the particle image velocimetry (PIV) is used to obtain the displacement field of five collapsed areas along the road.
This study is to obtain the horizontal displacement field of the collapsed and the surrounding area mainly based on the particle image velocimetry (PIV). Original aerial photographs took in 2007, 2008, 2010 and 2012 of five major collapses of Nantou 89th county road are collected and synthesized. ERDAS IMAGINE LPS modules are used to produce the DTM and orthophotos in order for later analyses. PIVview2C is used to conduct and process PIV analysis.
The position accuracy of the obtained orthophotos are ranging between 0.50 to 1.50 meters with the use of PIV analysis. The displacement accuracy deviation is between 0.14 to 4.59 m. However, PIV is using image characteristics similarity to conduct displacement vector calculation, the analyzed results may be affected by vegetation and shadow areas, e.g., movements at 32.5K. Therefore, the topographic variation analysis (DoD, Difference of DTM) and terrain sectional analysis are used to clarified and to double check if the analyzed results are reasonable. Through the calculation of the amount of collapsed movement and the displacement block via PIV analyses, the active moving zones of the five landslide locations are revealed. The failure processes and mechanisms are recognized. The results of this research can be useful for mitigation and design suggestion for the five collapsed places of the road in the future.

並列關鍵字

collapse ； PIV ； Difference of DTM ； ERDAS ； 89th county road

參考文獻

[11] 鮑鋒、苗君易、陳子良，「粒子影像測速技術及在流體測量上的應用」，航空、太空及民航學刊，第三十九卷，第二期，第093 - 098 頁，2007。

[16] 蔡維哲，「應用航空影像分析於邊坡位移場之方法學研究」，碩士論文，國立台灣大學土木工程學研究所，台北，2005。

[1] Slama, C.C. The Manual of Photogrammetry 4th. American Society for Photogrammetry,Falls Church, VA,1980

[5] 劉邦偉，「以特徵為基礎的方法對NASA JPL AIRSAR影像與航測正射影像進行」，碩士論文，國立成功大學測量工程學系，台南，2003。

[7] 詹景丞，「應用數值航空攝影測量於地形高程與地表特徵變異之評估」，碩士論文，國立中興大學土木工程學系，台中，2009。