Research And Implementation Of Barcode Image Recognition In Mobile Commerce

Mobile-Commerce utilizes mobile terminals'feature of access to the network at anytime and anywhere, so it makes users free from the restriction of Internet access to do e-business. M-commerce make the e-business be more effectively and easily execute. With the 3G network be commercially used, m-commerce development will be accelerated. Data collection is the main function of the mobile phone terminal. But this function doesn't work well because of the low capability of the mobile phone. And this hinders the m-commerce from development. Along with the development of the camera phone, the way of collecting data by recognizing the barcode image which captured by the camera on the phone is considered effective, economical and handful. But because of the low computing power, low memory power and low camera pixels, this way does not work well. The problems are such as low recognition capability, the recognition procedure take too much time. So, it's of significant meaning for m-commerce popularization to research on the technology of image recognition on average mobile phone which has low computing capability and low memory capability.This paper mainly discussed the image recognition under in circumstance of mobile phone. First, it analysis the typical mobile electronic commerce system's architecture and the main problems in this architecture(the problems are caused mainly by the contradiction between the demand of computing power, memory power in image recognition and the reality of low capability of the mobile phone). Then discussed the features and the code role of different kind of barcode and the differences between them, it's helpful to understand the differences between the barcode image recognition method in dealing with different barcodes. After that it did a lot of research in image processing algorithm and did a little improvement in the noise filtering and binarization process to made the algorithm more effective. The improved algorithm alleviate the contradiction between the demand of resources and the limited power of the mobile phone. At last this paper selected the code39 barcode and QR code to testify the algorithm. As QR code has the position detection pattern, so the recognition procedure made an improvement on the base of the procedure recommended in the GB/T18284-2000 National Standard to make it work more smoothly. According to the discussion on the code role of the barcodes, most part of the algorithm can also be used in other barcode image recognition procedure. Experiments were carried on the mobile phone of SonyEricsson K570 and made it clear that the algorithm can work well on the average camera mobile phone with low computing power and memory power. The result came up to the expectations.