| PDF417barcode, as one of the widely used two-dimensional barcodes, is beingapplied to various kinds of goods. There is almost no research on recognition ofPDF417barcode upon cylinder side surface, though goods in cylinder shape arecommon in life. To find a recognition method, features of PDF417barcode uponcylinder side surface have been analyzed. Aiming at providing possible solution torecognition of PDF417barcode upon cylinder side surface, researches on localization,rectification and decoding have been done. The main work and contributions are asbelow:1. To locate the PDF417barcode, an algorithm based on morphology and cornersdetection is proposed. To rectify and recognize PDF417barcode, we must separate thebarcode from the background. At first, morphology operations, throwing small areasand retaining large ones, have been done. Then corner detection is adopted to achievethe smallest area containing the barcode. Finally, localization of PDF417barcode uponcylinder side surface is achieved.2. To rectify the PDF417barcode, an algorithm based on local warp perspectivetransformation is proposed. To transform PDF417barcode upon cylinder side surfaceinto one upon plane, perspective distortion must be rectified. The PDF417barcode uponcylinder side surface is segmented into several sub areas first. Then perspectivetransformation will be done on these sub areas. Finally, these sub areas are concatenatedtogether and the rectified barcode is obtained.3. This paper completes the decoding process based on rectified barcode and therecognition system of PDF417barcode upon cylinder side surface. The rectifiedPDF417barcode is segmented into several symbol character columns, and the bar spaceratio is detected and calculated column by column. Then the codewords, obtained fromthe bar space ratios, are put together and decoded. At last, based on the processes oflocalization, rectification and decoding above, a recognition system for PDF417barcodeupon cylinder side surface is achieved. |
PDF Full Text Request