- 學位論文
影像檢測技術在晶片包裝紙帶沖壓製程的應用
Image detecting of paper carrier tape for chip-packing in stamping process
摘要
在晶片包裝紙帶沖壓生產線上，由於沖壓過程是高速地在進行，很 難直接檢測模具的磨損狀況，也不允許隨時停機檢視模具，所以必須由 製品來追溯判斷模具的狀況以確保線上產品品質。一般紙帶檢測是以人 工量測，不但量測速度慢又有視覺誤差，沖壓品質影響頗巨，本研究乃 是利用影像技術建構一台檢測 0402 晶片包裝紙帶之影像檢測系統，用來 檢測紙帶沖孔之尺寸與毛邊長度，以作為沖壓製品之品質檢驗及模具觀 測判斷與調整的依據，期望能建構優質快速的檢測系統以取代反覆又複 雜的人之視覺量測。 本研究之影像檢測系統係利用 CCD 擷取影像配合馬達帶動平台移 動，以一次量測 10 片紙帶，在影像處理中使用影像平移、臨界值法、輪 廓搜尋法及形態學等來做處理，10 片紙帶量測完畢可顯示其量測值、平 均值、3 個標準差(3σ)、最大/最小值、最大/最小之差值、綜合製程 能力指數(CPK)來確認產品之品質，並可由紙帶毛邊之長度與狀況去預 測模具是否需要調整或更換。實驗結果顯示此影像檢測系統量測紙帶每 片 64 秒鐘 (人工量測為每片約 7-8 分鐘) 可大量節省時問，最大變動 誤差為 0.005mm，誤差百分比最大為 3.86%，毛邊檢測最大誤差為 0.017mm，由毛邊檢測中可明顯發現紙帶毛邊位置並可量出其長度作為模 具調整或更換之時機，因此，此影像檢測系統可替代人工量測，降低人 為量測誤差及節省時間，以提高生產效能。
關鍵字
無資料並列摘要
The process of stamping proceeds is very fast in the production line of stamping for the paper carrier tape of chip. The measurement of the attrition of the mold couldn't be taken from the mold directly or stopped the machine for detecting When the machine is working. For the quality of products, the measurement most be taken from the product and then apply the data to find the attrition of the mold. In the convention, paper carrier tape of chip is checked artificially. However, it is slow and also has vision errors; therefore the pressing quality must be influenced. This research is using image technology to construct an image detection system for 0402 paper carrier tape of chip. The size of the paper carrier tape and the length of deckle edge could be measured by the system. With it we can get the system to adjust mold and the tool to check the quality of pressing products. With the detecting technique, a high quality and quick system to replace the traditional repeated complicated and artificial system can be constructed. The image detection system captures the image by CCD camera to cooperate with the motor to drive the platform moving, examine 10 slices of paper tape once. As dealing with image, we use image shift, threshold, Contour tracing and morphology. It can show the value of measure, average, 3 standard deviations (3σ), the maximum and minimum, subtract value of maximum and minimum, and process performance (CPK) to confirm the quality of pressing products. Also, the deckle edge condition can forecast to adjust the mold. Experimental result tells that it takes ninety second for the image detection system to measure paper carrier tape each one is 64 second, averagely nine second for one piece (artiflcial measure for each one is 7 to 8 minutes). It shows that we can save much time by the system. The maximum change error is 0.005 mm, the greatest percentage of error is 3.86% and the deckle edge detecting error is 0.0l7mm. In the detecting system to check deckle, we can find the position and the measuring length of deckle edge for adjusting or changing the mold suitably. Therefore, this image detection system can substitute artificial measure and reduce artificial mistakes and save time, for improving the producing efficiency.