| 摘要: | 橋梁會隨氣候及環境之交替而逐漸老舊劣化，故需定期巡檢，並籌編足額經費適時改善。我國橋梁約2.9萬座，並分由臺鐵局、高公局、公路總局及各縣市政府負責管養，按公路法相關規定，橋梁養護首重檢測，各橋梁管理機關除應適時針對所轄橋梁實施橋梁檢測作業外，並應針對損壞部分採取適當維修對策，以維行車安全。 依據「公路養護規範」、「公路鋼結構橋梁檢測及補強規範」及「公路鋼筋混凝土結構橋梁之檢測及補強規範」規定，我國橋梁檢測作業可分成「經常巡查」、「定期檢測」及「特別檢測」三類。其中「經常巡查」係於平時以乘車目視方式所實施之經常性檢查；「定期檢測」係透過步行、攀爬、作業車輛或船隻等方式，儘可能接近橋梁構件後，以目視判定橋梁狀況之定期性檢測。「特別檢測」係重大事故或災害發生後，為了解損傷程度而實施之不定期目視檢測。整體而言，橋梁檢測以目視為主，儀器為輔，故一般進行檢測作業時，多係以徒步及攀爬方式儘可能接近橋梁結構物後，再以目視判定橋梁狀況，如遇高橋墩或跨河橋梁時，雖可利用橋梁檢測車、高空作業車或小型船艇輔助趨近，然對於車流量大或偏遠山區橋梁而言，由於檢測車械體積較大、操作費時且成本較高，故相關巡檢作業仍有精進空間。交通部運研所於民國103年透過「橋梁檢測工具效能提升計畫」整合低價位無人飛行載具及相關軟硬設備，開發出有助於提升橋梁檢測作業品質及效率之旋翼式無人機橋梁檢測雛型模組。但因橋梁檢測現場環境多變(如障礙樹木、側風、GPS訊號不穩)，故尚有部分功能可再增進。 本研究提出防撞式UAV橋梁構件影像自動擷取系統，並著眼於四大功能模組之提升，包含:障礙物偵測模組(旋翼式無人機可偵測航行範圍內的障礙物並自動閃避)、定位精進模組(當旋翼式無人機於橋下時仍能具備定位能力)、飛行耐久模組(允許旋翼式無人機完成逾百米橋長的複合飛行路徑)、影像提升模組(搭載等級較高攝影機取得精細度更高的影像)和3D生成模組(將旋翼式無人機的檢測結果合併生成3D橋梁影像)。綜合而言，此防撞式UAV橋梁構件影像自動擷取系統將更能符合橋梁現地檢測之使用。同時，此系統具備與第二代橋梁管理資訊系統(TBMS2)整合之優勢，本研究預期能持續提升橋梁檢測作業之品質及效率。;There are 29,000 bridges in Taiwan. These bridges are managed by Taiwan Area National Freeway Bureau, Taiwan Railways Adminstration, and county governments. Based on several maintenance manuals regarding freeways, bridge inspection could be divided into general inspection, regular inspection and special inspection. For the general inspection, bridge managers should ensure the smooth of freeways. For regular inspection, bridge managers should observe the conditions of bridge components so that the safety of bridges could be confirmed. For the special inspection, bridge managers should check the conditions of bridges after some important hazards or events. In sum, bridge inspection depends on visual observation. When bridge managers investigate some bridges that cross rivers and mountains, they could apply specific vehicles during bridge inspection. However, these vehicles are big. Thus, the operation time and cost are high. In other words, the efficiency of bridge inspection should be improved. Associated with the rapid growth of information technology, various new equipment and techniques are also used in bridge inspection, such as small-size vehicles, aerial vehicles, 3D laser, and image recognition. For the aforementioned purpose, the previous project had applied Unmanned Aerial Vehicles (UAV) in bridge inspection successfully. However, the limitations still affected the usage of UAV. Therefore, this project offered four modules to increase the effectiveness of UAV for bridge inspection. These include obstacle avoidance, power enforcement, accurate position, and three-dimensional image. The obstacle avoidance module enables UAV to avoid sudden objects. The power enforcement could let UAV fly more than 30 minutes. Accurate position helps UAV to fly well under bridges. The three-dimensional image could combine the imagery of bridge elements. In sum, the efficiency of bridge inspection would be clearly increased.;研究期間：10405~10412 |
