- 學位論文
防護落石網之無線感測與影像告警系統開發
Development of Inclination Sensor System and Motion Detection System for Rockfall Protection Fence
摘要
台灣因獨特地理環境，每年經歷許多地震或颱風，在長年的影響下，使得山區邊坡的地質狀況變得脆弱且不穩定。隨著山區道路開發增加，鐵路等重要基礎設施沿山坡建置，落石災害之防護近十年來便受到大眾的重視。目前採取邊坡穩定性評估，再依潛在風險程度，安裝落石防護網被動攔截落石，而資源有限的條件下，少部份地區會額外加裝監測系統，如傾斜計裝於落石網的鐵柱上。在實務上，多直接採用精度高、價格高的傾斜計，適用於房物這類很細微的變動。然而對於受撞擊的落石網而言，角度變化比房物傾斜大上好幾倍，由此可見在需求面與供給面之間，存在著落差。 基於上述的現況，本研究旨在開發一套監控系統，包括偵測鐵柱傾斜程度及分析影像。為了在功能與設備成本之間取得平衡，本研究整合現有模組，如ESP32、MPU6050、Raspberry pi和LoRa，ESP32特有的傳輸協定ESP-NOW，使資料傳輸輕量化，LoRa則用於長距離告警系統。而影像分析利用開源軟體 Motion，並透過現地試驗，檢視其實用性。期許落石監測系統得以普及，並易於客制化設定以應對不同場址。
並列摘要
Taiwan is located on the boundary of tectonic plates and in subtropical climate region that induce earthquakes and typhoons yearly. Since the slopes become much fragile, it is getting common to implement disaster prevention system on unstable slope. However, applying disaster prevention is time-consuming and costly. In this regard, we proposed a configurable and affordable monitoring system installed on rockfall protection fence to inspect rockfall event when taking place. The system mainly consists of two parts, inclination sensors system and motion detection system. We integrate inertial measurement unit, MPU6050, and microcontroller, ESP32, as inclination sensor. It is characterized with lightweight transmission protocol that make data delivery efficient. Besides, we adopt LoRa module to trigger distant alarm system. The motion detection system is comprised of Raspberry pi and IP camera. Once it obtains the motion with changed pixels more than threshold, it will capture the certain frames and notify relevant people. The prototype of the monitoring system is test on the site on Fanglan road.