基于改進雙色模型的視程障礙類天氣現象識別
DOI:
作者:
作者單位:

作者簡介:

通訊作者:

中圖分類號:

基金項目:

公益性行業(氣象)專項天氣現象自動化觀測技術研究(GYHY200906032)資助


Recognition of Weather Phenomena with Low Visibility on Basis of Improved Dichromatic Atmospheric Scattering Model
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 圖/表
  • |
  • 訪問統計
  • |
  • 參考文獻
  • |
  • 相似文獻
  • |
  • 引證文獻
  • |
  • 資源附件
  • |
  • 文章評論
    摘要:

    針對各種視程障礙類天氣現象,在現有雙色大氣散射模型的基礎上,通過對大氣散射光方向的修正、對矢量夾角算法的改進以及加入了濕度參數信息,提出了一種改進的雙色模型算法,用于識別視程類天氣現象。改進算法模型考慮了大氣散射光方向的不完全一致性以及景物色度像素值在RGB空間中的三維性,因而具有更高的穩定性。該算法嵌入到視程障礙類天氣現象識別系統中并計算出相應的特征值,最后通過經驗閾值識別天氣現象;通過對算法的改進并對哥倫比亞大學一個季度和甘肅張掖觀象臺2個月的數據進行處理,結果顯示該算法對于霧、霾、揚沙、浮塵及沙塵暴等視程障礙類天氣現象具有良好的判別效果,對上述兩個地區的總體識別率分別為732%、905%

    Abstract:

    Concerning various kinds of low visibility weather phenomena and based on the existing dichromatic atmospheric scattering model, a modified calculating method of dichromatic model is proposed, to range weather phenomenon via revising the atmospheric scattering direction, improving the calculating method of vector angles and adding humidity parameter information The improved algorithm model has higher stability, considering that the scatter light direction is not all consistent and pixels have three dimensional character in the RGB space Embedding the algorithm into the model and computing the corresponding eigenvalues, these weather phenomena can be recognized through the empirical thresholds By improving the calculating method and processing the seasonal data of Columbia University and two month data of Zhangye Observatory, it is shown that the improved method is excellent in identifying various kinds of low visibility weather phenomena, such as fog, haze, dust blowing, dust floating, and sand storm, whose total identification rates for the above two regions are 732% , 905%, respectively.

    參考文獻
    相似文獻
    引證文獻
引用本文

郭佳,楊玲,吳可軍,馬舒慶,胡樹貞.基于改進雙色模型的視程障礙類天氣現象識別[J].氣象科技,2013,41(2):236~241

復制
分享
文章指標
  • 點擊次數:
  • 下載次數:
  • HTML閱讀次數:
  • 引用次數:
歷史
  • 收稿日期:2011-12-31
  • 定稿日期:2012-09-07
  • 錄用日期:
  • 在線發布日期: 2013-04-11
  • 出版日期:
您是第位訪問者
技術支持:北京勤云科技發展有限公司
午夜欧美大片免费观看,欧美激情综合五月色丁香,亚洲日本在线视频观看,午夜精品福利在线
>