基于dsp图像采集和苹果颜色分级

《基于dsp图像采集和苹果颜色分级》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、摘要苹粟分缀是孳累商品化处理鞠棱心环节。袋羽机器视觉技术对苹果大小、形状、色泽、缺陷等强麓晶质遂{i全覆分缓，可鼗代，洚绫熬夫工分缀释撬电癸级方式。鑫予毽臻Pe掇避孬莩鬃蘑态分缎的缺点，而DSP轻CPLD褶结合的视颟黼像采集髓理技术已经报成熟，所戳本文迸{7蕊予DSP的苹果动态分级系统研究。零文结合苹聚羚疑魏窭嚣嚣甏，搭建其有巍额篆集嚣嚣示、审滩逶蔼、计数端秘、数享赣如、激太瓣接西餐翡能瓣DSP匿像赡理开麓缀。在就基础上采强DSP／BIOS疆露戆梅，并发撬行雾任务，窭褒葶豢强慷靛采集、筵濮耧显示。其中，DSP／BIOS麴藏

2、键帮努—VP日瓣动糕枣设计，采用丁层次清晰的类，微型驱动模型，使或用程序与硬件无关。运过毙较H1S颧篷搂墼帮聪尉豢爨摸型，袁接装零嚣者鼢熬蓑整建立攀巢壤蕊燕泽菠数字模懿和着色痿数枣模型。首先搽闱u分鬣进行澎色图像背最分渤，得到苹巢像素；接蒋联合苹果像豢的v分爨秘U分最，计爨U+V-128，得到勉、黄、绿三种苹果果面熬魏的取值范隧，即雕建立苹栗颜魏色泽度数字模型；媛后，红、黄、绿三神苹果象黼灏色像素数与苹果像素韵比德，霹灸苹采凝惫藩雹发数字模型。懑过天王势级帮搬器援觉努级豹缝莱魄较，薅卷其毒趣葑戆一羧德。将苹果定位、图像采集

3、处理和苹果弹出功能协调一致，寞现按苹果直径大小的动态分级。为夸嚣苹集影搜、擞色秘嫒臻舅法鹣秘态努缓奠嶷了基穑。美镞蠲；TMS320DM642，戡Iv，荦栗灏色，瓣豫莱集，韵态势缎AbstractApplegradingis￡kcoreofRscommercialprocessing，Theapplicationofmachinevision敞hologyingradingapplesaccordingtotheirappearanceindexsuchassize，shape，color,defect，elc·，callt

4、aketheplaceoftraditionalhumangradingandmechanic·electronicgrading．Takingintoconsiderationthedisadvantagesofapplereal—timegradingonPCandthematurityofvideocap境u'e-and-processtechniquescombinedwithDSPandCPLD,thepresentpaperresearchesoilapplereal-timegradingsystembase

5、donDSEGroundedontherealneedofapplegradingSystem，DSPimageprocessboardisdevelopedwithsuchfunctions§svideocapturinganddisplayingserialcc日axmunication，counterport}digitaloutput，andEthernetinterface。DSP／BIOSstrucRLreisthenimplementedwithmulti-tasktorealizeappleimagecap

6、turing，processing，anddisplaying．VPDDKdesign，whichisthekeyofDSP／BIOS，adoptsclearclass／minidrivermodeltomaketheapplicationprogramsirrelevantfromthehardwaIe．Furthermore,theapplecolorandcolorationdegreemodelsare赋HpbycomparingHISand磷Ⅳcolormodelsandusingthelatterchroma．

7、Firstly,thecolorimageisdivided魄theUofYUV,andtheappleimageisderived．Secondly,U+V-129iscalculatedwiththatoftheapplepixels，whichresultsillthedatascopeofthered,yellow,andgreenapplesrespectively,Thecolordegreemodelissetupinthisway．Finally,氆eapplecolorationmodelisobtain

8、edthrought赫ratiooftheapplefacecolorplxelBumandthatofthethreetypesofapples．Theresearchprovestl-mtmachinevisiontechnologygradingresultnmtcheswellwiththatt