2015-01-302024-05-14https://scholars.lib.ntu.edu.tw/handle/123456789/657460摘要：中央氣象局氣象測站現行盛行能見度以及天氣現象觀測是經由觀測員目測並佐以能見度目標圖加以記錄，電子式能見度儀(Visibility meter)則透過水滴或氣膠對於陽光散射程度來推估能見度，兩者在氣象觀測的比對鮮見討論。中央氣象局基於天氣觀測守視技術的更新與自動化，已在幾年前開啟台灣地區部分測站裝置Mobotix M12戶外型數位攝影機與能見度儀之觀測測試。本研究延續第一年計畫資料蒐集(中央氣象局現有八套戶外型數位攝影機與六套能見度儀器之影像樣本和能見度儀數據蒐集，以及測站鄰近的環保署PM2.5濃度逐時資料，來與觀測員目視觀測進行交叉分析)，以及第二年計畫Mobotix M12、VM95以及BlueEyes等影像分析的能見度閾值調整之成果，提出第三年工作計畫來總結完成「影像分析在天氣觀測之應用」目標；具體工作項目包括 (1) 根據2014年8~10月M12資料持續調整各處測站VB程式的最佳化index value。 (2) 於配置M12攝影機之測站電腦安裝VB程式，蒐集人工觀測以及能見儀二種資料，進行交叉比對。(3)持續蒐集台北測站BlueEyes環場攝影機影像資料，來進階完成台北測站不同方位數位能見度守視效果。最後，透過各測站其他氣象觀測要素(比如太陽輻射量、絕對溼度、向風速等)，來改進各測站能見度與天氣現象自動觀測效果，協助中央氣象局達成天氣觀測自動化輔助之技術開發典範。<br> Abstract: The prevailing visibity and weather phenomenon observation at CWB weather stations are recoreded by observabors’ eye. The principle of electronic visibility meter takes the solar scattering charactertics of aerosol & rain drop to estimate the visibility distance. The intercomparison between these two observation methods is rare globally. In order to improve the automated process of weather observation at CWB weather stations, CWB has installed several Mobotix M12 digital videos and visibility meters for operation planning test. In this third-year stage, we extend our experiences on data accumulated from the first yeart and index value for image processing threshold from M12, VM95 and BlueEyes videos in second-year stage, several working items will be approached to finish the original objective of this program: (1) based on the data collected in August to October of 2014, Turning our VB image filter package for better threshold pf index value (2) install this VB package in each weather station for continuous recod and compare it with the record from observers’ eye and visibility meter. (3) keep the measurement of BlueEyes circle-view video on the roof of Taipei weather station, to study the performance of wide-view video image. Finally, the other parameters (such as solar radiation, humidity and wind filed) from the regular observayion are studied to know their impact on the digital iamge analysis in weather watch.能見度天氣現象影像分析weather phenomenonvisibilityimage analysis