基于Hough变换的高分辨率遥感影像城市直线道路提(5)

2009.3　　　　　　　　　　　　　　　　　　　　　遥感信息综述　　　　　　　　　　　　　　　　　　　EuSAR[C]12006.

AReviewofInversionoftheForestHeightby

PolarimetricInterferometricSAR

LITing2wei,LIANGDian2nong,ZHUJu2bo

(SchoolofElectronicScienceandEngineering,NationalUniversityofDefenseTechnology,410073)

Abstract:Inversionofvegetationparametersbypolarimetricincreasinginterestinthefieldofradarremotesensingandmanymethodsofhavebeenproposed.Inthispaper,thesemethodsareclassifiedtofourclassesthesemethodsaresystematicallystudied.After2wards,theestimatedandtheadvantagesanddisadvantagesofthesemethodsarepointedout.ofofvegetationparametersareabstracted.Attheend,thefutureresearchfieldsareproposed.

Keywords:polarimetricSAR;interferometricSAR;inversionofvegetationparameters

(上接第18页)

6　结束语

本文根据遥感影像中城市道路的特性,先设定

阈值进行图像分割,得到包含道路网轮廓的二值图像,使用Hough变换检测道路所在直线,对直线进

行道路判断,再将所得道路段进行修剪、连接形成道路网,实现道路提取。实验结果表明,使用Hough变换检测道路直线时,能减少影像中的汽车及建筑物阴影对道路提取的干扰。该方法能有效的从高分辨遥感影像中提取城市直线道路。

参考文献

1　史文中,朱长青,王昱1从遥感影像提取道路特征的方法综述与展望[J]1测绘学报,2001,30(3):257～2611

2　朱长青,王耀革,马秋禾,史文中1基于形态分割的高分辨率遥感影像道路提取[J]1测绘学报,2004,33(4):347～35113　王珂娜,邹北骥,钟志华,等1基于Hough变换的焊接接头自动识别技术研究[J]1湖南大学学报,2004,31(5):99～10314　瞿洋,杨利平1Hough变换OCR图象倾斜矫正方法[J]1中国图象图形学报,2001,6(2):178～18115　文贡坚,王润生1从航空遥感图像中自动提取主要道路[J]1软件学报,2000,11(7):957～9641

UrbanStraightRoadExtractionfromHigh2ResolutionRemote

SensingImageBasedonHoughTransform

LVJian2gang,WEIChun2tao

(DepartmentofCivilEngineering,GuilinUniversityofTechnology,Guilin541004)

Abstract:Accordingtotheurbanstraightroadcharacterofhigh2resolutionremotesensingimage,thispaperpresentsanapproachtoex2tractroadnetwork.Basedontheroadnetworkoutlinefromimagesegmentation,houghtransformisusedtodetectthestraightlines,whichcontaintheroadinformation.Judgingthestraightlinewhetheritisroadornot,andgettheroadsections.Aftertrimmingandconnectiontheroadsectionstheroadnetworkisformed,thustheroadextractioniscompleted.Experimentre2sultsshowthatthisapproachisefficientforextractingurbanstraightroadfromhigh2resolutionremotesensingimage.Keywords:houghtransform;roadextraction;remotesensingimage;imagesegmentation

91