Study And Implementation Of QR Code Image Recognition Based On DSP

As one kind of two-dimensional barcode, QR code has many advantages such as large information capacity, high reliability, fast decoding, omnidirectional decoding and expressing Chinese more effectively. It's a common code of two-dimensional barcode for mobile phone. With the development of 3G, QR code will have broad prospects.Most studies of QR recognition system are based on PC recently. These systems are inconvenient and need to improve the recognition accuracy and speed on images with uneven illumination and complicated background. AQR code recognition algorithm for such images is designed and implemented in this paper. A QR code recognition system with image collection, recognition, result output functions is also designed and implemented based on TI-TMS320DM642 DSP. This paper is made up of the following parts:The existing recognition technologies, QR code structure and encoding theories are studied and analyzed for the design of the algorithm and the system.The QR code image recognition algorithm, which is composed of pretreatment, location, sampling and decoding algorithm, is designed and implemented in this paper after analyzing many algorithms. Local threshold method, improved position exploration pattern method and rotated sampling network' are used for binarization, location and sampling respectively.An embedded QR code recognition system is designed and implemented. The hardware design includes the selection of processor and the design of system structure, the software design includes the module designs of image collection, image recognition, result output and system boot.MATLAB simulation results and system tests indicate that the embedded QR code recognition system based on DM642 can recognize and decode QR code images with arbitrary rotation under slightly complicated background. Good efforts on recognition accuracy and efficiency are achieved by the system.