QR码图像识别的关键技术

The Key Technology of QR Code Image Recognition

摘要: 本文针对利用普通拍照设备在自然环境下抓拍到的含QR码的非标准图像，首先对其进行灰度化处理，其次采用中值滤波、全局阈值二值化、图像矫正等方法与过程完成图像的预处理工作，得到了比较标准的QR码图像。在图像矫正中，采用了Sobel算子边缘提取、Hough变换、图像旋转、平面投影转换及双线性插值法等方法。最后对于矫正后的图像，经过确定寻像图形、确定版本和格式信息、恢复掩膜、恢复码字并纠错、数据码字译码等一系列的处理，对其实现QR码译码。

Abstract: This paper aims at the nonstandard image of QR code, captured in a natural environment using ordinary photo equipments. First of all the image is grayed; secondly using median filter, the global threshold binarization and image correction, we complete image preprocessing and get more standard QR code image. In the process of image correction, this paper adopts Sobel operator edge detection, Hough transform, image rotation, plane projection transformation, bilinear interpolation method and so on. Finally, for the corrected image, the paper completes the QR code decoding by a series of processes of determining the finder patterns, the version and format information, restoring mask film, restoring code words and correcting error, data decoding and so on.