基于自适应采样的全息图像压缩感知降噪方法

打开文本图片集

Holographic image compressive sensing noise reduction based on adaptive sampling

LI Bingliang'，CHENBo¹，QINMengwei¹，YANG Jing² （1.CollegeofElectricalEngineering，NorthChina UniversityofScienceand Technology，TangshanO63210，China; 2.College ofContinuing Education，NorthChina UniversityofScience and Technology，Tangshan O63210，China）

Abstract：During theresearch of digital hologramsamplingandreconstruction，thenoiseofholographicimage willffect therecostructionresult.Theexistingcompresiveholographyfailstotake intoaccountthesparsityandtexture feature distributionofdigitalholographicimages，andthesamplingrateisalocatedunreasonably.Inviewoftheabove，adigital hologramnoisereductionmethodbasedonadaptiveblockcompressivesensingisproposed.Inthemethod，theintensityoflight fieldisusedasthemeasureofimagetexturedetails.Onthebasisofthepre-samplinginformation，theholographicimageis dividedintodiferentinformation blocksandtheadaptivesamplingrateiscalculated.Theadaptiveunder-sampledreconstruction ofthenoisyholographicimageisrealizedbycompresivesensing.Theresultsofthenumericalsimulationexperimentsshowthat theproposedmethodcanimprovetheSNR（signal-to-noiseratio）andsimilarityofholographicimagesefectivelyinthecasethat thesamplingrateexceedsO.4whenitisappiedtodiferentreconstructionalgorithms.Itcanbeseenthattheproposedmethod canmakefulluseofthesparsityandtexturefeaturesofthedigitalholographicimages，andallcatethesamplingrateofdierent noisesubblocksreasonablyincomparisonwith the traditionalreconstructionalgorithms.Therefore，theproposedmethodcan efectively improve the quality of the object image reconstructed by optical reconstruction.

Keywors：digital holography;blockcompresivesnsing;noisereduction;edgeinformation;image reconstruction;adaptive sampling

0 引言

数字全息术采用光电传感器记录全息图，将复杂的光学重现方式转换为数字计算问题。（剩余7936字）