在Lightning接頭的生產製程,製程中仍存有許多不穩定因素,半成品電線容易於焊接加工的前後產生瑕疵。在焊接加工前,可能產生極性錯誤、絕緣體剝除不完全、連錫等瑕疵;在焊接加工後,印刷電路板上可能產生連焊、焊件偏移等瑕疵,而目前的品質檢測方式是以人工透過儀器放大監看,並手動排除瑕疵品,故本研究針對Lightning接頭的生產瑕疵,開發專用的AOI瑕疵檢測系統。 本研究於生產Lighting接頭的焊接設備,架設兩組機器視覺檢測站,以PC端連接兩站的相機,取得待測物影像,並以Visual C# 搭配Halcon撰寫檢測軟體,自動判別良品與不良品,而人機介面顯示其檢測結果,若檢測為不良品,則透過Arduino觸發設備控制器來取消後續製程。 檢測結果顯示,深圳版本的前檢測(焊接前)辨識時間每組約為0.44秒、檢出率為96.4%;台灣版本的前檢測(焊接前)辨識時間每組約為0.24秒、檢出率為97.8%,後檢測(焊接後)辨識時間每組約為0.19秒、檢出率為100%,而觸發設備控制器使後續製程不進行加工的時間約為1秒。
Lots of unstable factors still occur in the producing process of the lightning connector, easily happen if the semi-finished wires are getting soldering before and after. It may cause the defects on the wire, such as wrong polarity, incomplete plastic insulation or in bridge etc., before soldering. It may makes the problems like solder bridge or component shifted on the Printed Circuit Board (PCB) after soldering as well. Moreover, all quality inspection implement by the manual instrument inspection and remove defect by human. According to the above, this paper developed the AOI system for these issues. Two sets of machine vision inspection stations are set up on the pre-soldering and post-soldering equipment of the lighting connector. The PC side connects with two sides of cameras and captures the images of cores, identifies the defects automatically through the inspection software coded by Visual C # and Halcon image processing library. If the inspected results show defects, they will cancel the after process through the Arduino UNO. The before inspection results show that the average time of the identification in Shenzhen version is 0.44 seconds per set, the inspection rate is 96.4%. The average time of identification in Taiwan version is 0.24 seconds per set, the inspection rate is 97.8%. At the after inspection in Taiwan version, the average time of identification is 0.19 seconds per set, the inspection rate is 100%. And it’ll need about 1 second to cancel the after process if the set point is opened.