運用影像處理於航照影像之自動河道變遷分析

Automatic River Course Change Analysis Using Image Processing in Aerial Photographic Images

黃帥豪(Shuai-Hao Huang)

指導教授 ： 張元翔

中原大學/工學院/資訊工程研究所/碩士(2008年)

https://doi.org/10.6840/CYCU.2008.00407

摘要

民國88年間之集集大地震與其後連年颱風所挾帶之豪雨，造成地層嚴重鬆動、鬆散之土石隨著流水沖刷造成大量的土石堆積，使得大甲溪的樣貌有大幅的改變。本研究主要目的是憑藉航空照片影像，以影像處理技術發展一套”自動河道變遷分析系統”，此系統可以用來偵測航空照片中的河道影像，觀察2001年(桃芝颱風後)、2004年(敏督利颱風後)以及2006年(艾莉颱風、海棠颱風後)的航空照片並分析經過這些天災肆虐之後，從德基壩至馬鞍壩之間大甲溪主要河道於不同時段的變遷情形。本系統設計主要分成前置處理，並藉由影像處理方式找到主要河道，於影像中定義河道位置，及自動化評估河道變遷等資訊。系統效能由系統計算結果藉由與人工方式作相互比較得到，本研究所辨識的個別時期河道初期辨識率約為88%，其辨識結果可建立出一個多時段大甲溪主要河道變遷分析系統。總體而言，本系統除了可有效記錄河道變遷資訊，並可協助地質人員簡化河道變遷分析工作。若未來新的航空照片輸入，可利用本系統獲得最新的河道辨識，並和過去的辨識結果做比較，期能對其變遷狀況進行監測及分析，達到鑑往知來，作為災害防治的參考，以防止進一步的災害發生。

關鍵字

河道變遷 ； 影像處理 ； 自動辨識 ； 航照

並列摘要

The Chi-Chi earthquake and the following series of typhoon events in years which brought heavy rainfalls caused the rock layers and masses strongly disturbed, and huge amount of sediments coming from the landslides, debris flows, and floods has made dramatically change of the Dar-Chia river course along recent years. The objective of this research was to develop an “Automatic River Course Change Analysis (ARCCA) system”, that could be used to detect aerial photographic images and to determine if the image include the main river course. The system was used to analyze changes of river course between the De-Ji Reservoir and the Ma-An dam during the years 2001(After the Toraji Typhoon), 2004(After the Mindulle Typhoon), and 2006(After the Aere and Haitang Typhoon). Our system design included pre-processing and image processing techniques that were used to locate and define main river course, to automatically analyze the information with respect to river course changes. Our preliminary results demonstrated a river course detection of approximately 88% by evaluating the system performance based on manual results by geologists. These results could be used to achieve the Dar-Chia river analysis system among various time-frames. In summary, our system could help effectively record the relevant information, and geologists simplify the labor-intensive task of river change analysis. Ultimately, if new aerial-photographic images are available, the system could be used to acquire river channel detections for both present and past results easily, and to further understand river channel changes quantitatively, thus achieving the goal of the prevention of future disasters.

並列關鍵字

Aerial Photography ； Automatic Interpretation ； River Course Change ； Image Processing

參考文獻

[5] 王景平、林銘郎、鄭富書、游明芳、周坤賢，松鶴地區土石流災害歷史之探討, Journal of Chinese Soil and Water Conservation, vol.36, no. 2, pp. 203-213, 2005.

[7] 楊適修，山崩教室，國立中央大學應用地質研究所，

[17] 鄭錦桐、張玉粦、邱顯晉、林彥享、顧承宇、許世孟，大甲溪上游台電水力發電設施在強震豪雨後之衝擊及因應對策芻議，中興工程，工程技術，第九十五期，頁21-30，2007。

[1]行政院農委會水土保持局全球資訊網

[6] 劉守恆，衛星影像於崩塌地自動分類組合之研究，國立成功大學 地球科學研究所，碩士論文，2002。