Research On Location And Recognition Algorithms Of Two-Dimensional Barcode Based On Mobile Phones

The barcode technology has been developing rapidly since 20^th century and has been used in various trades. It is wide applications of one-dimensional barcodes that speed up the data collection and information processing, and it provides the great contribution for the modernization and socialization of administration. Due to the limit of the information capacity, one-dimensional barcode is just a label for goods. They must be used with database. With the development of high technology, it is pressed for using barcode to express more information within the limited printing space, which could realize the description for goods. Two-dimensional barcodes appeared just about to meet upwards requirement. Two-dimensional barcodes have many types, such as QR, PDF417 and Maxicode.Nowadays, people's demands for mobile phones are not merely confined to conversation itself. They pay much close attention to integrating more and more applications. Usually, the devices utilized for barcode detection are using laser beams. Since many mobile devices are already equipped with digital cameras, they can incorporate this kind of application as well. The purpose of the thesis is to locate, detect and recognize two-dimensional barcode through mobile phones.Barcode detection is an important step. Before its recognition, barcode needs to be located by filtering text and other signs in the image. In this thesis an algorithm is presented that locates quick response code from images captured by camera phones. The algorithm uses DCT-transform properties to enhance quick response area in order to distinguish bar code from other areas and morphological operations to smooth the detected quick response area. Experimental results show that this method has good performance and the correct localization from images in different skews. After localization, an algorithm of barcode recognition is proposed in this thesis. Barcode recognition is a matter of edge detection, which has long been done through the maximum value of the first derivatives or zero crossing of the second derivatives. When the barcode is of higher density, adjacent edges interact each other due to the influence of the point spread function of image collection system. Therefore, this kind of method becomes invalid. This thesis analyzes point spread function according to the structure of QR code, then restores original image. Experiments show that the algorithm presented in this thesis is able to effectively overcome blurring affects caused by the point spread function and apparently promote the recognition rate, and the algorithm is better than that based on edge detection.