欢迎来到天天文库
浏览记录
ID:36784875
大小:4.89 MB
页数:66页
时间:2019-05-15
《基于边缘特征的匹配算法改进研究》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、颂.1:论义某于边缘特征的匹配算法改进研究摘要针对图像匹配全局搜索影响匹配速度的问题,本文引入了遗传算法,将其与Hausdom距离相结合来优化图像匹配的搜索过程。首先对基准图像和实时图像进行预处理,以自适应选取阂值的Canny算子来提取图像的边缘作为特征单元;然后引入遗传算法对待配准图像进行匹配操作;在遗传操作中利用改进的STMHausdorff足巨离来构造遗传算法的适应度函数,以此通过遗传算法确定出最优的变换参数,完成基准图像和实时图像的匹配;最后通过实验仿真验证了算法的有效性。Hausdor艘巨离作为一种评价两个图像位置关系的量化标准,以其很强的抗干扰和容错能力
2、被广泛应用于图像匹配之中。然而单纯的Hausdorff足巨离对噪声和孤立点均比较敏感,导致误匹配率较高。各类改进的Hausdorff足巨离形式能在特定的匹配环境中克服这些不足,但面对本课题中复杂的成像畸变和复合噪声的情况,仍不能获得理想的匹配效果。本文采取了一种先对标准方差的Hausdorfi腽巨离(sTMHD)进行排序,尔后再求其部分均值的Hausdo硪距离改进形式。这种经改进的ST心D能较好的克服噪声、伪边缘及部分遮挡对匹配精度和稳定性的影响,在配准速度和精度上取得较为理想的效果。关键词:图像匹配,边缘,c锄ly算子,Hausdor唧巨离,遗传算法Abstrac
3、t颂I:论文Globalsearchfortheimagematchingthespeedoftheimpactofmatchingproblems,thispaper,weintroduceageneticalgorithm,withtheHausdorffdistancewillbeacombinationofimagematchingtooptimizethesearchprocess.First,thereferenceimagesandreal—timeimagepre—processing,inordertoselectadaptivethreshold
4、Cannyoperatortoextracttheedgeimageasafeatureunit.Andthendealwiththeintroductionofgeneticalgorithmtomatchtheimageregistrationoperation.GeneticmanipulationintheuseofstandarddeviationmodifiedHausdorffdistanceimprovedgeneticalgorithmtoconstructthefitnessfunction,geneticalgorithmtodetermine
5、theoptimaltransformationparameters,thecompletionofthereferenceimagesandreal-timeimagematching.Finally,throughexperimentsimulationtheeffectivenessofthealgorithm.●Hausdorffdistancebetweentwoimagesasanevaluationoftherelationshipbetweenthelocationofthequantitativecriteria,、加thitsstronganti
6、—interferenceabilityandfaulttolerancearewidelyusedinimagematching.Hausdorffdistance,howeversimpleonthenoiseandisolatedpointsaremoresensitive,resultinginahigherrateoffalsematches.VarioustypesofimprovementintheformofHausdorffdistancematchingaspecificenvironmenttoovercomethesedeficiencies
7、,butinthefaceofthecomplexityoftheissueofdistortionoftheimagingnoiseandcomplexsituation,theyCallnotachieveanidealeffectmatch.TlliSarticletakesafirststandarddeviationmodifiedHausdorffdistance(STMHD)sort,fortheirpartofSeoulafterthemeanimprovementsintheformofHausdorffdistance.STMHDthisim
此文档下载收益归作者所有