基于fpga的雾天图像恢复算法研究

基于fpga的雾天图像恢复算法研究

ID:23282520

大小:2.87 MB

页数:56页

时间:2018-11-06

基于fpga的雾天图像恢复算法研究_第1页
基于fpga的雾天图像恢复算法研究_第2页
基于fpga的雾天图像恢复算法研究_第3页
基于fpga的雾天图像恢复算法研究_第4页
基于fpga的雾天图像恢复算法研究_第5页
资源描述:

《基于fpga的雾天图像恢复算法研究》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、ABSTRACTTheeffectivenessofoutdoorvisionsystemsaregreatlylimitedinfoggyconditionsforthevideoandimagescapturedbyimagingdevicedegradesbecauseoftheatmospherescattering.Real-timerecoveryforthefoggingvideoandimageshasimportantpracticalsignificance.Thisthesisstudiedthefoggyimagedegradationmodelbasedo

2、nanalysisoffoggingimagedegradationmechanism.Aneffectivedefoggingalgorithmwithlowcomplexityisproposed.Thealgorithmachievedreal-timeclarificationforvideoandimagescombinedwithparallelprocessingapproachandthe"pipeline"structureofFPGA.Themaincontentsareasfollows:1)Thedegradationmechanismoffoggyimag

3、eandtheatmosphericscatteringmodelwerein-depthanalyzed.Advantagesanddisadvantagesoftwoimagede-hazingalgorithmsbasedonatmosphericscatteringwereexpoundedafterdiscussedindetail.Alloftheseworkmadefulltheoreticalpreparationsforfollowingstudy.2)Fordarkchannelprioralgorithmhasmanyshortcomingssuchasexi

4、stinghaloswhenprocessingskyarea,highcomplexityandhighresourceconsumptionwhenimplementedontheFPGA.Afastdarkchannelpriordefoggingalgorithmisproposed.Firstofall,thephenomenonofhaloinskyareawassolvedandalsothehardwareresourceconsumptionwasreducedbysettingthresholdvaluetoatmosphericlightA.Thenoptim

5、izedsolvingmethodsoftransmission,transformedtheserialcomputingofthetransmissionandatmosphericlighttoparallelcomputing.FinallyaccordingtothecircuitdesigncharacteristicsoftheFPGA,introducedguidedfilteringinsteadofsoftmattingwhichcomplexityisunrelatedtothewindowsize,thusimprovedtheefficiencyofalg

6、orithm.3)Asetofimagede-hazingsystembasedonDE2-70boardashardwareplatformwasdevelopedinthisthesis.Thehardwarecircuitsoffastdarkchannelprioralgorithmweredesignedafterdesigningofoverallarchitecture,thenfulfilledthetransplantofthealgorithm’sFPGAplatform.Someimportantpartofthemodules,suchasthedarkch

7、annelmodule,atmosphericlightmoduleandDehazemoduleweresimulated.Finallyachievedreal-timefoggyimagerecovery.4)Theresultsofthealgorithm’sMATLABsimulationshow,thefastdarkIII万方数据channelprioralgorithmproposedinthisarticleprocessesquickly,andt

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。