Abstract: The QR code acquired by camera always comes with some distortion, so it needs to be recognised to the standard QR code before decode. Aimming at the QR coderecognition, distortion and correction is analyzed and studied in this paper. Some inevitable distortion still existed based on the tilt correction and geometric correction；the traditional method can’t sample the QR code accurately. According to the problem, this paper proposes the adaptive match method, acquire the effective sampling region of QRcode by the matching rate of two adjacent pixel row（column）. Experiment shows that the method is real-time with good stability, it can sampling the QR code fast and accurately.

Key words: QR code, recognition, tilt correction, geometric correction, sampling, adaptive match

摘要： 通常由相机获取的QR码图像都带有一些失真，所以在译码前需要对获取的QR码图像进行识别以得到标准规格的QR码。针对QR码识别中的失真和校正进行了分析研究，解决了某些QR码经过倾斜校正和几何校正后仍存在一些无法避免的失真而无法被传统方法准确取样的问题，提出了一种自适应匹配取样法，根据相邻行（列）像素的匹配度准确获取QR码的模块有效取样区域。实验证明该方法稳定性好，能够快速准确地对QR码进行取样。

关键词: 快速响应码, 识别, 倾斜校正, 几何校正, 取样, 自适应匹配