無人飛行載具於露天礦場邊坡監測之應用

Abstract

現今我國礦業開發以露天開採的大理石及矽砂為大宗，而露天礦場最主要面臨之災害為落石及崩塌問題，常造成生命與財產損失。為達到以低成本方式對此類災害防範預警之目標，本研究嘗試利用無人飛行載具(Unmanned Aerial Vehicle, UAV)針對礦場邊坡進行其穩定性監測。 位於苗栗縣南庄鄉的桃源興業礦場於2021年1月曾發生東南側邊坡崩塌事故，因此本研究選定該礦場為研究場址，於2021年10月開始，使用無人飛行載具於不同時期拍攝全礦及邊坡。截至2022年11月共進行六次飛行，取得五期影像，並藉由Agisoft Metashape Pro商用套裝軟體提供之運動恢復結構 (Structure from Motion, SfM)技術建置礦場3D點雲及數值高程模型(Digital Elevation Model, DEM)，再以ESRI ArcGIS軟體進行不同期影像比對。 根據2022年3月取得之影像與前期比較，本研究發現礦場西北側邊坡有滑動趨勢，配合同年6月進行現地調查，發現該處邊坡出現冠部裂隙、崩崖、蝕溝等山崩前兆現象，隨即將此監測資訊提供礦場業者以進行處置，成功防止崩塌災害再次發生。經處置後，同年10月、11月再度取得兩期影像並進行比對，發現該礦場西北側邊坡目前已無先前大面積滑動之跡象。 本研究結果顯示利用無人飛行載具針對露天礦場邊坡進行監測，或可有效預警礦場邊坡之崩塌災害。由於現今無人飛行載具取得容易，且相關軟體的使用門檻亦不高，將其運用於防範礦場邊坡崩塌災害上，具有很大的應用價值。

Nowadays, the mining industry in our country primarily involves the open-pit extraction of marble and silica sand. The main hazards faced by open-pit mines are rockfalls and landslides, which often result in loss of life and property. To achieve the goal of low-cost disaster prevention and early warning for such hazards, this study attempts to utilize Unmanned Aerial Vehicles (UAVs) to monitor the stability of mine slopes. The Taoyuan Xingye mine, located in Nanzhuang Township, Miaoli County, experienced a landslide on its southeastern slope in January 2021. Therefore, this study selected this mine as the research site. Starting from October 2021, UAVs were used to capture images of the entire mine and its slopes at different periods. By November 2022, six flights were conducted, obtaining five sets of images. These images were processed using the Structure from Motion (SfM) technique provided by Agisoft Metashape Pro to construct multi-period 3D point clouds and Digital Surface Models (DSMs) of the mine. The images from different periods were then compared using ESRI ArcGIS. Comparing images obtained in March 2022 with those from earlier periods, this study found a sliding trend on the northwestern slope of the mine. Combined with an on-site investigation conducted in June of the same year, precursory landslide phenomena such as crown cracks and erosion gullies were observed on the slope. This monitoring information was immediately provided to the mine operators for appropriate action, successfully preventing another landslide disaster. After remediation, two more sets of images were obtained in October and November of the same year for comparison, revealing that the northwestern slope of the mine no longer showed signs of large-scale sliding as observed previously. The results of this study demonstrate that using UAVs to monitor open-pit mine slopes can effectively provide early warnings for landslide hazards. Given that UAVs are now easily obtainable and the related software is relatively easy to use, their application in preventing mine slope landslide disasters holds significant value.