中国农村水利水电
中國農村水利水電
중국농촌수이수전
CHINA RURAL WATER AND HYDROPOWER
2015年
8期
69-73
,共5页
面向对象%分级提取%线性水系%LBV变换
麵嚮對象%分級提取%線性水繫%LBV變換
면향대상%분급제취%선성수계%LBV변환
object-oriented classification%hierarchical extraction%linear river system%LBV transformation
以获取干旱半干旱区精准水域分布信息为目标，利用水域的水文特征，采用水系分级化处理方法，分级识别水域信息从而开展基于环境减灾卫星 HJ的水环境遥感识别监测研究。首先对影像进行预处理，构建环境卫星的归一化差异综合水体指数，利用阈值法粗区分水体影像中的面域水体和细小线性水体。构建分离背景条件分离山体阴影、植被、荒漠、裸地等大部分背景信息；然后对影像进行 L BV变化增强；结合 L BV特征与粗提取结果，面向对象分割进行精细化提取面状水体和细小水系；最后利用区域生长算法修复水系断流现象获得最终完整的水体分布情况。结果表明，该方法能够准确识别面域和细小支流水体边界，显著减少植被和阴影对水系边界的影响，基本消除背景值对水体识别的干扰，水系识别完整度为91．3％，用户精度为92％。比常用水体提取方法提取效果好，且同样适用于其他水系。
以穫取榦旱半榦旱區精準水域分佈信息為目標，利用水域的水文特徵，採用水繫分級化處理方法，分級識彆水域信息從而開展基于環境減災衛星 HJ的水環境遙感識彆鑑測研究。首先對影像進行預處理，構建環境衛星的歸一化差異綜閤水體指數，利用閾值法粗區分水體影像中的麵域水體和細小線性水體。構建分離揹景條件分離山體陰影、植被、荒漠、裸地等大部分揹景信息；然後對影像進行 L BV變化增彊；結閤 L BV特徵與粗提取結果，麵嚮對象分割進行精細化提取麵狀水體和細小水繫；最後利用區域生長算法脩複水繫斷流現象穫得最終完整的水體分佈情況。結果錶明，該方法能夠準確識彆麵域和細小支流水體邊界，顯著減少植被和陰影對水繫邊界的影響，基本消除揹景值對水體識彆的榦擾，水繫識彆完整度為91．3％，用戶精度為92％。比常用水體提取方法提取效果好，且同樣適用于其他水繫。
이획취간한반간한구정준수역분포신식위목표，이용수역적수문특정，채용수계분급화처리방법，분급식별수역신식종이개전기우배경감재위성 HJ적수배경요감식별감측연구。수선대영상진행예처리，구건배경위성적귀일화차이종합수체지수，이용역치법조구분수체영상중적면역수체화세소선성수체。구건분리배경조건분리산체음영、식피、황막、라지등대부분배경신식；연후대영상진행 L BV변화증강；결합 L BV특정여조제취결과，면향대상분할진행정세화제취면상수체화세소수계；최후이용구역생장산법수복수계단류현상획득최종완정적수체분포정황。결과표명，해방법능구준학식별면역화세소지류수체변계，현저감소식피화음영대수계변계적영향，기본소제배경치대수체식별적간우，수계식별완정도위91．3％，용호정도위92％。비상용수체제취방법제취효과호，차동양괄용우기타수계。
In order to obtain accurate water distribution information in arid and semi-arid areas ,utilize hydrological characteristics of waters and method of hierarchical hydrography ,and carry out the Remote Sensing Monitoring of Huanjing (HJ) ,First ,the image that is founding the normalized water index extraction of HJ is pre-processd and threshold segmentation is used to differentiate be‐tween water region and narrow water branch .The conditions to separate background information are constructed :hill shade ,vegeta‐tion ,desert ,bare land and so on .Then the images with LBV characteristics are enhanced .LBV features are combined with the crude extraction results .In order to solve the problem of cutoff areas in the river ,the last step is repairing hydrographic net along the growth direction by applying regional growth algorithm .Results show that this method makes full use of hierarchical division of the extraction ,can identify the border of water region and the narrow river accurately and reduce the effects of shadows and other vegetations from border of river system significantly .And it can eliminate the interference of the water identification from back‐ground value ,and have a high integrity of water identify .The completeness of rivers is 91 .3% ,the user accuracy is 92% .This method is better than common water extraction and can also apply to other water systems .
