Summary: | 碩士 === 大同大學 === 資訊工程學系(所) === 93 === 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.
|