應用UAV影像分析技術於固床工變動及周圍流場之研究

Application of UAV Imaging Technology for Ground Sill Variation and Flow Field Analysis

賴進松(Jihn-Sung Lai) ； 韓仁毓(Jen-Yu Han) ； 李豐佐(Fong-Zuo Lee) ； 張文鎰(Wen-Yi Chang) ； 楊淑媛(Shu-Yuan Yang)

《農業工程學報》 64卷2期 (2018/06) Pp. 1-11

https://doi.org/10.29974/JTAE.201806_64(2).0001

摘要

近年來氣候變動造成降雨強度增加，降雨強度增加若伴隨強降雨發生機率也跟著加大，則強降雨所產生之洪水可能造成橋梁損壞、堤防溢堤或堤岸潰堤等災害，進而影響交通及堤內人身安全。固床工為國內穩定河床坡降及保護其上游構造物之常見工程，規劃良好之固床工可防止河川改變流路、穩定河道底床坡降及岸壁淘刷等功效，達到保護上游橋梁或是水工構造物之安全。濁水溪中游國道一號中沙大橋下游側之固床工為本研究區域，本研究應用無人飛行載具、影像邊緣偵測並配合標的物紋理偵測進行固床工之自動化辨識技術，以及大尺度粒子影像分析法，進行固床工在颱風豪雨衝擊後之變動判識以及固床工周圍之流場變化分析。本研究所研發之固床工自動化辨識技術能準確判識試驗室固床工變化，而且能辨識80%以上之固床工範圍，輔以人工辨識後則可有效率地辨識出固床工面積變化。此外，本研究應用大尺度粒子影像分析法，利用現地UAV所拍攝影像進行固床工周圍之流場變化分析，而試驗室測試結果顯示大尺度粒子影像分析法能分析出正確流場，並計算出正確流速，於現地量測結果亦顯示大尺度粒子影像分析法能分析出固床工周圍之正確流場。

關鍵字

固床工 ； 無人飛行載具 ； 自動化辨識技術 ； 大尺度粒子影像分析法

並列摘要

In recent years, climate change is a major factor to increase rainfall intensity and frequency. The increased rainfall intensity and frequency will also increase the probability of flash flood with abundant sediment transport. The floods caused by heavy rainfall may cause damages to the bridge, embankment, hydraulic works and the other disasters. Those damages will affect traffic, transportation, human safety and property. Ground sill is a major engineering method to stabilize the river bed and protect its upstream structures. Well-designed ground sill can prevent the flow path changing and river bank scouring flush. In addition, ground sill can stabilize the river bed slope and protect upstream bridges or hydraulic structures. The study area is located at Zhongsha Bridge downstream of National Freeway No. 1 in the middle of the Zhuoshui River basin. In this study, we adapt Unmanned Aerial Vehicle (UAV) to take images in the field for ground sill variation and flow field analysis. The image edge detection and texture detection are presented to carry out the automatic identification technology on ground sill variation. In addition, Large Scale Particle Image Velocimetry (LSPIV) is developed to analyze flow field around ground sills. Study event concentrates on the typhoon and heavy rainfall effects. This study developed the image edge detection and the texture detection that can accurately identify the ground sill changing in laboratory condition. The accuracy can reach more than 80% on the variation of ground sills in laboratory condition. However, supplemented by manual interpretation is needed in the field to effectively improve and identify the area changing of ground sill. This study also uses LSPIV method and UAV images on the flow field analysis around ground sill in the field. The results show that LSPIV method can analyze the flow field and flow velocity correctly in laboratory tests. In the field, the presented LSPIV method can successfully analyze the flow field around ground sills using UAV images.

並列關鍵字

Ground sill ； Unmanned Aerial Vehicle ； Automatic identification technique ； Large Scale Particle Image Velocimetry