摘要
對於大部份的產業而言，條碼系統已是不可或缺。然而，在條碼的使用上，仍然還有一些缺點限制其自動化的運作。例如，在使用雷射讀碼機時，必須手動操作，並且得相當靠近條碼，才能夠讀取。此外，條碼也需要是相當平坦的才行，否則就得拉平它，就像一般便利商店店員經常需要做的事。 為排除這些手動操作，我們發展一個可以藉由相機或攝影機作為擷取裝置與電腦化處理的條碼系統。透過影像擷取與切割下來的條碼，是我們想要辨識的目標樣本。然而，這些條碼影像有可能是不夠清晰、被扭曲、或處於摺傷的狀況，都將影響電腦的判讀。因此，我們提出一套能夠處理扭曲變形之條碼的扭曲條碼辨識系統。首先，利用叢集法對條碼影像進行二元化的處理，接著藉由條碼影像的分析程序記錄下條碼品質的資訊。根據這些資料，我們就可以正確的讀取出這些有缺陷的條碼影像。
並列摘要
Nowadays, barcode systems have been so popularized for most industries. However, there are still some disadvantages limiting their automation, such as the laser reader has to be put close to the barcode manually and the barcode needs to be pretty flat. For avoiding manual operation, we want to develop a camera-based barcode decoding system. The barcode which was cut from a captured image is our sample. However, the barcode may be a unclear, twisted, or folded one. To conquer these problems, we present a twisted barcode recognition system. The handling processes start from the binarization the barcode image by clustering method. Then, we examine the quality of the binarized barcode image and keep track for assisting the decision of the decoding policies. From our experimental result, we can recognite most of the defective barcode image even it is demaged quite severely.