欢迎来到天天文库
浏览记录
ID:36832997
大小:3.41 MB
页数:81页
时间:2019-05-16
《基于神经网络的图像识别与分类技术及应用研究》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、华中科技大学硕士学位论文基于毒审经网络的图像识别与分类技术及应用研究(摘鬻)【关键强】模式识别襻经鄹终特链提取Bp算法一、、f蘩予糖缀褥络黢甏缘识聚与分类技零是随慧姿翦}}算裁按术、鬻像缝理、大王餐熊、模式汉剿毽谂等发鼹起米懿一秘凝型妥缘谈裂技术,悬京传统翦墨像识嬲方法熬基秣上融会襻缀耀络算法的一耱嬲缀识别方法,对于当蘸静交遗管理、王妲生产、医疗诊敷等领域麴发矮霄鬟大攘县意义。本文结含汽车牌照臼动识别袈绞取繁镪袭麟缺陷识别与分类系绞鼹个实例藿点分缨了纂予{孛经网络的图像识嬲与分必技术的原理及应月方法。j/本文锋对缨像议男
2、《技零孛足耱常魁麴识别方法懿不是,楚攀介缨了襻经鼷络算法原理季曩一弛事卓经鼹络模型⋯..蓖馕燃络3、程较为困难,阻碍其在图像识别系统审更好鼢癍箱,葭既本文搬稻实践经验重赢介绍了对BP神经黼络的网络缭构、擎露簿注等翡改进方法,荠结合汽举牌照裔动识别系统和带钢表面缺陷识剃与分类系统瓣令安瓣详缨介绍了麓改避静BP耐络设计鹩分类器的实现过程。本文同辩瞧寝热餐疆离BP享牵缀秘络泛他能力方面镞了一定髂探讨。,,n9”L本文提畿懿罄予棒经两络的图像识剐与分类技术己成功地成用在汽车牌照臼麓谖躺系统秘带镑表嚣缺陷谖掰与分类系统中,并飘效采蠹著,同甜该方法对圈豫谖聚系统森其襄领域孛麓瘦弱瞧蠢一定魏指导意义膏7,华中科技大学硕士学位论文Th4、eResearchonTechnology&ApplicationofImageRecognition&ClassificationBasedonANN(ABSTRACT)[keywords]PatternRecognition,ANN,FeatureExtraction,BPAlgorithm强etechnologyofimagerecognitio珏andclassificationbasedonANNiSakindofnew—styleimagerecognitiontechnology,whichisdevelo5、pedduetothetheoryofpresentcomputertechnology,imageprocession,artificialintelligenceandpaRernrecognition,andproducesimportantdirectivemeansonthedevelopmentofcurrenttrafficmanagement,industryproductionandmedicaldiagnose.Thisarticleemphasesonthetheoryandrealizationo6、fimagerecognitionandclassificationbasedonANK。onthebaseofthesystemofautomaticrecognitionofcarlicenseplateandsurfacedefectrecognitionandclassificationofsteelstrip.Thisarticle.aimingatafewkindsoflimitationofrecognitionmethodcommonlyusedinimagerecognition,conciselyin7、troducestheprincipleofANNalgorithmandakindofcommonANNmodel-BP,accordinglyputsforwardakindofmethodofimagerecognitionanddetailedlyreferstotherealizationofthewholeprojectofimagerecognitionandclassificationbasedonANNandthedescriptionofeveryfunctionalmoduleastheexampl8、eofthesystemofautomaticrecognitionofacarlicenseplateandsurfacedefectrecognitionandclassificationofsteelstrip;italsoindicatesthatfeatureextractionandfeaturesele
3、程较为困难,阻碍其在图像识别系统审更好鼢癍箱,葭既本文搬稻实践经验重赢介绍了对BP神经黼络的网络缭构、擎露簿注等翡改进方法,荠结合汽举牌照裔动识别系统和带钢表面缺陷识剃与分类系统瓣令安瓣详缨介绍了麓改避静BP耐络设计鹩分类器的实现过程。本文同辩瞧寝热餐疆离BP享牵缀秘络泛他能力方面镞了一定髂探讨。,,n9”L本文提畿懿罄予棒经两络的图像识剐与分类技术己成功地成用在汽车牌照臼麓谖躺系统秘带镑表嚣缺陷谖掰与分类系统中,并飘效采蠹著,同甜该方法对圈豫谖聚系统森其襄领域孛麓瘦弱瞧蠢一定魏指导意义膏7,华中科技大学硕士学位论文Th
4、eResearchonTechnology&ApplicationofImageRecognition&ClassificationBasedonANN(ABSTRACT)[keywords]PatternRecognition,ANN,FeatureExtraction,BPAlgorithm强etechnologyofimagerecognitio珏andclassificationbasedonANNiSakindofnew—styleimagerecognitiontechnology,whichisdevelo
5、pedduetothetheoryofpresentcomputertechnology,imageprocession,artificialintelligenceandpaRernrecognition,andproducesimportantdirectivemeansonthedevelopmentofcurrenttrafficmanagement,industryproductionandmedicaldiagnose.Thisarticleemphasesonthetheoryandrealizationo
6、fimagerecognitionandclassificationbasedonANK。onthebaseofthesystemofautomaticrecognitionofcarlicenseplateandsurfacedefectrecognitionandclassificationofsteelstrip.Thisarticle.aimingatafewkindsoflimitationofrecognitionmethodcommonlyusedinimagerecognition,conciselyin
7、troducestheprincipleofANNalgorithmandakindofcommonANNmodel-BP,accordinglyputsforwardakindofmethodofimagerecognitionanddetailedlyreferstotherealizationofthewholeprojectofimagerecognitionandclassificationbasedonANNandthedescriptionofeveryfunctionalmoduleastheexampl
8、eofthesystemofautomaticrecognitionofacarlicenseplateandsurfacedefectrecognitionandclassificationofsteelstrip;italsoindicatesthatfeatureextractionandfeaturesele
此文档下载收益归作者所有