公路图像拼接的算法研究

公路图像拼接的算法研究

ID:43750946

大小:915.85 KB

页数:69页

时间:2019-10-13

公路图像拼接的算法研究_第1页
公路图像拼接的算法研究_第2页
公路图像拼接的算法研究_第3页
公路图像拼接的算法研究_第4页
公路图像拼接的算法研究_第5页
资源描述:

《公路图像拼接的算法研究》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库

1、河北工业大学硕士学位论文公路图像拼接的算法研究姓名:孙荣艳申请学位级别:硕士专业:模式识别与智能系统指导教师:顾军华20081101公路图像拼接的算法研究ALGORITHMRESEARCHOFTHEHIGHWAYIMAGEMOSAICABSTRACTImagemosaicisatechnologythatcarriesonthespatialmatchingtoaseriesofimageswhichareoverlappedwitheachother,andfinallybuildsaseamlessandhighqualityimagewhichhashighresolutionandb

2、igeyeshotcomparedwithasingleimage.Imagemosaicisanimportantresearchingfieldofimageprocessing,andhaswidelyapplicationsinthefieldsofphotogrammetry,computervision,remotesensingimageprocessing,medicalimageanalysis,computergraphicandsoon.Becausethecharacteroftheroadimagesiscomplexandunobvious,thisarticle

3、hasbeengivenaspecialmethodofimagemosaic.Thethesisfirstintroducesthepresentresearchstatusofthetechnologyofimagemosaic,thushasdemonstratedthebroadapplicationprospectofthetechnologyofimagemosaic.Thenthethesisdiscussesthebasicstepsofimagemosaic,imageacquisitionmethodsandmodelsofimagemosaic.Moreover,itr

4、esearcheswithemphasisinthetwocoretechnologyofimagemosaic,whicharethetechnologyoftheimageregistrationandtheimagefusion.Theexistingmethodsofimageregistrationandimagefusionareintroduced,andtheirperformanceaswellastheexistingproblemsisanalyzed・Accordingtothecharacteristicsofroadimage,theroadimageswould

5、bedividedintodamagedandundamagedimages.Eachtypeisgivenadifferentmethodofsplicing.Theimagesofdamagedroadscontainsalotoffeatures:First,imagesarestratifiedbytheGaussianpyramid;second,theneachlevelisdetectedfeaturepointsbytheHarrisalgorithm;third,thefeaturepointsaredividedbytheFCMclustering;finally,the

6、imagesaremosaicbycorrelationcoefficientandarefusedbytheWeightedcoefficients・Theundamagedimagescontainfewcharacteristicinformation.First,imagesaredetectededgesbythealgorithmofSUSAN;second,theimagesareenhancededgeinformationandremovedthenoisebythemorphologymethods;third,theimagesareblockedbytheedges;

7、forth,theimagesaredetectedthefeaturesthroughtheartificialfishalgorithm;finally,theimagesarestitchedbycorrelationcoefficientandarefusedbytheWeightedcoefficients・Bothalgorithmstakefulladvantageoftheinterestpo

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。