Small Parts Classification with Flexible Machine Vision and a Hybrid Classifier

2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP) Pub Date : 2018-11-01 DOI:10.1109/M2VIP.2018.8600819

Keyur D. Joshi, B. Surgenor

{"title":"Small Parts Classification with Flexible Machine Vision and a Hybrid Classifier","authors":"Keyur D. Joshi, B. Surgenor","doi":"10.1109/M2VIP.2018.8600819","DOIUrl":null,"url":null,"abstract":"A Flexible Machine Vision (FMV) Inspection System has been developed that requires minimal retuning in hardware and software as applications are changed up. The flexibility of the system was evaluated by applying it to an inspection problem with three different types of small parts: plastic gears, plastic connectors and metallic coins, with minimal retuning when moving from one application to the others. The system was required to differentiate between 4 different known styles of each part plus one unknown style, for a total of 5 classes. In previous work, a hybrid Support Vector Machine (SVM) classifier was developed for the connector application. When applied to the coin application, the hybrid SVM could not achieve the target performance of 95% accuracy. A new hybrid that method that combines SVM and an Artificial Neural Network (ANN) or ANN-SVM classifier was subsequently developed to overcome this problem and the results are presented in this paper. The image library used in this study is available at http://my.me.queensu.ca/People/Surgenor/Laboratory/Database.html","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"6","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/M2VIP.2018.8600819","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 6

Abstract

A Flexible Machine Vision (FMV) Inspection System has been developed that requires minimal retuning in hardware and software as applications are changed up. The flexibility of the system was evaluated by applying it to an inspection problem with three different types of small parts: plastic gears, plastic connectors and metallic coins, with minimal retuning when moving from one application to the others. The system was required to differentiate between 4 different known styles of each part plus one unknown style, for a total of 5 classes. In previous work, a hybrid Support Vector Machine (SVM) classifier was developed for the connector application. When applied to the coin application, the hybrid SVM could not achieve the target performance of 95% accuracy. A new hybrid that method that combines SVM and an Artificial Neural Network (ANN) or ANN-SVM classifier was subsequently developed to overcome this problem and the results are presented in this paper. The image library used in this study is available at http://my.me.queensu.ca/People/Surgenor/Laboratory/Database.html

查看原文本刊更多论文

基于柔性机器视觉和混合分类器的小部件分类

开发了一种灵活的机器视觉(FMV)检测系统，当应用程序发生变化时，需要最小的硬件和软件返回。通过将该系统应用于三种不同类型的小部件(塑料齿轮、塑料连接器和金属硬币)的检测问题，评估了该系统的灵活性，并且在从一种应用程序移动到另一种应用程序时返回最小。系统需要区分每个零件的4种不同的已知风格和1种未知风格，共5类。在之前的工作中，针对连接器应用开发了一种混合支持向量机分类器。当应用于硬币应用时，混合支持向量机无法达到95%准确率的目标性能。为了克服这一问题，本文提出了一种将支持向量机与人工神经网络(ANN)或ANN-SVM分类器相结合的混合方法。本研究使用的图像库可在http://my.me.queensu.ca/People/Surgenor/Laboratory/Database.html上获得

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)

自引率

0.00%

发文量