儀表視覺檢測系統之研發

Development and Implementation of Phantom Identification Technology for Meter

賴逸仁(Yi-Ren Lai)

指導教授 ： 許源鏞

國立虎尾科技大學/工程學院/機械與機電工程研究所/碩士(2011年)

摘要

近年來，車用儀表朝數位化與CAN-Bus通訊網路發展。數位式車用儀表大量生產時，每一儀表組及其零組件皆有特定規格，需做進料與產品檢測，非常的耗費人力與時間，且產品的檢測瑕疵的認定，往往會因檢測人員的標準不一而有所不同，對產品的品質造成影響。為使產品檢測更省時、省人力，並使檢測標準一致，需要進行儀表視覺檢測系統開發。
本文藉由免費OpenCV影像處理軟體，進行儀表視覺影像辨識技術開發，包括指針角度辨識、燈號顯示位置等；並運用Visual C++ MFC視窗介面軟體，建立儀表視覺檢測系統。為因應各型號儀表不同檢測需求，需建立檢測數據資料庫，以及產生儀表檢測所需之頻率、電壓、電阻或數位I/O訊號，並可儲存及列印檢測資料。另配合不同檢測需求，能做各種儀表的教導學習功能，以使檢測系統具有通用性檢測功能。
實驗結果顯示，本文開發完成之儀表視覺檢測系統，導入生產線進行產品檢測，能快速並精確的完成產品檢測的工作，避免人為檢測瑕疵等問題，達到節省人力時間與檢測標準一致性的要求，且能大大的提高產品品質及經濟效益，值得業者投入開發。

關鍵字

數位式車用儀表 ； 視覺影像辨識系統 ； 數位訊號產生模組

並列摘要

n recent years, the vehicle faced digitizing and the controller area network development with the measuring appliance. In mass production, every automobile instrument and associated parts and components have specific specifications. Product inspection is necessary in feed processing and production and it requires large number of laborers and significant amount of time. However, standards of defect inspection often vary from one inspector to another. As a result, the quality of products is affected. In order to realize efficient inspection in both the number of laborers and the amount of time, a unified standard for inspection is needed.
This article utilizes the OpenCV phantom processing software, carries on the measuring appliance vision phantom identification technology, including the indicator angle identification, the lamp signal, etc. We also utilize Visual C++ the MFC Windows interface software, establish the measuring appliance vision examination system, use advanced visual image recognition system as the inspection platform for products. The digital signal module generates signals with specific resistance, voltage and frequency for automobile digital instruments and the system enables visual inspection instead of traditional inspection performed by laborers.
With the combination of human machine interface, visual image recognition algorithm, and the digital signal module, the visual image recognition system can perform efficient and accurate product inspection by lowering labor cost and avoiding human errors in inspection. This system can be widely used in modern automated production line where only a small number or laborers are needed.

並列關鍵字

automobile digital instruments ； visual image recognition system ； digital signal module

參考文獻

[4] 許源鏞、康振祥、吳建男、柯嘉城，“車輛區域控制網路系統整合與驗證”，國立虎尾科技大學學報，Vol. 27，No. 3，pp.13-20，經濟部學界科專計畫96-EC-17-A-16-S1-051-A1，2008。

[5] 吳建男，具CAN-Bus功能車用儀表之研發，國立虎尾科技大學碩士論文，台灣雲林(2009)。

[18] Gonzalez, Rafael C.; Woods, Richard E.,Digital Image Processing, 2nd Edition, Prentice Hall, 2002.

[29] Gonzalez, Rafael C.; Woods, Richard E.,Digital Image Processing, 2nd Edition, Prentice Hall (2002).

[30] Cheng-Chien Liu, Digital image processing-Introduction to Remote Sensing, NCKU, Nov. 18, 2004.