ASIA unversity:Item 310904400/112470
English  |  正體中文  |  简体中文  |  Items with full text/Total items : 90452/105769 (86%)
Visitors : 11946777      Online Users : 404
RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version


    Please use this identifier to cite or link to this item: http://asiair.asia.edu.tw/ir/handle/310904400/112470


    Title: 不同卷積神經網路應用在自動光學檢測的比較
    Comparison of Different Convolutional Neural Networks for Automated Optical Inspection
    Authors: 呂嘉銘
    LU, CHIA-MING
    Contributors: 資訊工程學系
    Keywords: 自動光學檢查、卷積神經網路、深度學習
    automated optical inspection、Convolutional Neural Network、deep leaening
    Date: 2019
    Issue Date: 2019-11-11 11:35:02 (UTC+8)
    Publisher: 亞洲大學
    Abstract: 國內有很多傳統產業工廠的生產線,運用到自動光學檢測(Automated Optical Inspection,簡稱AOI),AOI運用光學原理和機械視覺的技術來偵測產線上的物件是否有缺陷或瑕疵,能代替以人力使用光學儀器速度慢、容易誤判的缺點,以取代目前之人工目視檢測作業,改善目前人工檢測所不足的檢測精度,並提昇檢測速度與減少誤判率,最終目標達到減少人工成本與維持產品品質的理想目標,但AOI並沒有辦法分類瑕疵,只能判斷有無瑕疵。
    近年來,深度學習的快速發展,許多深度學習模型在影像分類上都有成功的應用,將深度學習與AOI做結合,運用圖像分類的方法來分類AOI影像瑕疵,以補足傳統AOI的缺點,分辨物件的瑕疵能對不同的瑕疵進行修補修復的處理,使得產能進一步提高。本研究主要重點在於使用不同卷積神經網路模型對自動光學檢測影像進行模型訓練和預測,來判讀瑕疵的分類,藉以提升透過數據科學來加強 AOI 判讀之效能,並使用其預測的準確度比較不同模型的優劣。
    There are many production lines of traditional industrial factories in our country, which are applied to automated optical inspection (AOI). AOI uses optical principle and mechanical vision technology to detect whether the objects on the production line are defective or flawed. It can replace the shortcomings of slow speed and easy misjudgment that using optical instruments by human, and replace the current manual visual inspection operation, improving the detection accuracy of the current manual detection, improving the detection speed and reducing the false positive rate. The ultimate goal is to achieve the ideal goal of reducing labor costs and maintaining product quality. However, AOI has no way to classify it, only to judge whether there is any flaw.
    In recent years,due to the rapid development of deep learning, many deep learning models have been successfully applied in image classification.We combine deep learning with AOI and use image classification to classify AOI images to compensate for the shortcomings of traditional AOI. Disadvantages, the handling of the parts can be repaired and repaired by different defects, which further increases the production capacity. Identifying defects in objects and repairing different defects, further increasing productivity. The main focus of this study is to use different convolutional neural network models to model and predict automatic optical detection images to determine the classification of defects. Enhance the effectiveness of AOI interpretation through data science and compare the pros and cons of different models with the accuracy of their predictions.
    Appears in Collections:[Department of Computer Science and Information Engineering] Theses & dissertations

    Files in This Item:

    File Description SizeFormat
    index.html0KbHTML18View/Open


    All items in ASIAIR are protected by copyright, with all rights reserved.


    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback