由於快速回應碼(QR code)能儲存比傳統一維條碼更多的資訊內容,可透過手機鏡頭輕鬆取得文字資訊,因此已被廣泛使用,但在瓶罐上的QR碼卻因曲面變形影像常無法被正確解讀,因此也限縮了QR碼可使用的範圍,因此本論文提出適用於瓶罐曲面上QR碼的影像校正技術,首先進行影像擷取處理以找出物體的直徑參數,其中包含影像模糊化、邊緣檢測、膨脹、補洞及侵蝕等運算,再經由平面展開數學模型校正變形的曲面QR碼影像,最後使用內插法對展開影像所擴增的像素計算其灰階值,完成曲面變形QR碼修復校正的目的,解決QR碼無法應用在瓶罐等彎曲面上的問題。實驗中以10度為單位增加彎曲角度,結果顯示原影像彎曲角度超過150度時QR碼將無法辨識,經本文所提方法展平校正後,使用QR碼解碼軟體則能完成正確解碼,證明本文應用影像展平數學模式的方法確可有效改善曲面QR碼無法被解碼的問題。
A quick response code (QR code) is a two-dimensional bar code that contains not only numbers but also characters. QR codes are now used in both commercial tracking applications and convenience-oriented applications aimed at mobile phone users. A defect has been found that QR codes cannot be correct decoded due to the curved surface of bottles. In this thesis, an image adjusting technique is proposed to improve the decoding correctness of a distorted QR code image. The processes include image segmentation, expansion and interpolation. Series of experiments using varied bending angles of the QR code have been tried. The QR codes can be correct decoded for bending angles up to 150 degree. Experiment results demonstrate the excellence of the QR code image adjusting technique.