计算机应用
計算機應用
계산궤응용
COMPUTER APPLICATION
2015年
z1期
130-134
,共5页
路志英%尹静%程亮%刘海%贾惠珍
路誌英%尹靜%程亮%劉海%賈惠珍
로지영%윤정%정량%류해%가혜진
列车效应%识别%跟踪%预报%Hu矩
列車效應%識彆%跟蹤%預報%Hu矩
열차효응%식별%근종%예보%Hu구
train effect%identification%tracking%forecast%Hu moment
提出了一种对“列车效应”进行识别和预报的新方法,该方法首先从“列车效应”历史样本的雷达反射率图中寻找整体带状回波的规律,提取单体回波的特征,并利用基于粗糙集理论的数据挖掘技术寻找隐含在单体回波特征中的有用知识,从而建立“列车效应”识别客观模型。然后利用该模型分别从整体和局部的角度对预处理后的实时天气状况进行识别。对于识别出的“列车效应”,利用整体带状回波的移动距离和旋转角度以及单体回波的面积变化和Hu矩进行跟踪。最后,利用相邻时刻“列车效应”的相关数据,建立云团族谱关系,通过云团的运动惯性对“列车效应”进行外推预报。实验结果表明:“列车效应”正确识别命中率是82.77%,能准确地预报出“列车效应”在6 min和12 min后的情况。该方法科学有效,有助于减轻“列车效应”灾害天气造成的损失。
提齣瞭一種對“列車效應”進行識彆和預報的新方法,該方法首先從“列車效應”歷史樣本的雷達反射率圖中尋找整體帶狀迴波的規律,提取單體迴波的特徵,併利用基于粗糙集理論的數據挖掘技術尋找隱含在單體迴波特徵中的有用知識,從而建立“列車效應”識彆客觀模型。然後利用該模型分彆從整體和跼部的角度對預處理後的實時天氣狀況進行識彆。對于識彆齣的“列車效應”,利用整體帶狀迴波的移動距離和鏇轉角度以及單體迴波的麵積變化和Hu矩進行跟蹤。最後,利用相鄰時刻“列車效應”的相關數據,建立雲糰族譜關繫,通過雲糰的運動慣性對“列車效應”進行外推預報。實驗結果錶明:“列車效應”正確識彆命中率是82.77%,能準確地預報齣“列車效應”在6 min和12 min後的情況。該方法科學有效,有助于減輕“列車效應”災害天氣造成的損失。
제출료일충대“열차효응”진행식별화예보적신방법,해방법수선종“열차효응”역사양본적뢰체반사솔도중심조정체대상회파적규률,제취단체회파적특정,병이용기우조조집이론적수거알굴기술심조은함재단체회파특정중적유용지식,종이건립“열차효응”식별객관모형。연후이용해모형분별종정체화국부적각도대예처리후적실시천기상황진행식별。대우식별출적“열차효응”,이용정체대상회파적이동거리화선전각도이급단체회파적면적변화화Hu구진행근종。최후,이용상린시각“열차효응”적상관수거,건립운단족보관계,통과운단적운동관성대“열차효응”진행외추예보。실험결과표명:“열차효응”정학식별명중솔시82.77%,능준학지예보출“열차효응”재6 min화12 min후적정황。해방법과학유효,유조우감경“열차효응”재해천기조성적손실。
A new identification and prediction method of“train effect” was studied. Firstly, rules of the whole band echo were found and characteristics of monomer echo were extracted from radar reflectivity graph, and then implicit useful knowledge in the monomer echo characteristics was mined by data mining technology based on rough set theory. Consequently, an objective identification model of “train effect” was established. Secondly, preprocessed real-time weather was identified separately in global perspective and local perspective by the identification model. Thirdly, if the result of recognition was“train effect”, displacement distance and rotation angle of whole band echo as well as area change and Hu moment of monomer echo were used to track the “train effect”. Finally, cloud genealogy relationship was established by relevant data of “train effect” in adjacent moment. By cloud genealogy relationship and inertia theorem the "train effect" was forecasted. The test results show that the accuracy of recognition rate is 87. 33%, and the “train effect” after six minutes and twelve minutes can be accurately forecasted. The method is scientific and effective to recognize and predict“train effect”, and is helpful to reduce the loss caused by“train effect”.jiangyilan1023@163. com)