QR码作为一种可高速、全方位识别的二维条码,已经在各行各业中得到了应用,并具有很大发展潜力.本文针对QR码图像的识别过程中可能遇到的图像噪声干扰、QR码占实际拍摄图像的一部分、倾斜、几何失真等问题进行了研究,提出了通过改进型中值滤波去除噪声,使用QR码符号特性定位并对倾斜图像进行旋转,对失真不严重的图像采用控制点变换和双线性插值方法进行几何校正.实验证明, 本文的方法简单有效, 可以用于提高条码的正确识别率.-As a high-speed, full identification of the two-dimensional bar code, QR code has been applied in all walks of life, and also has great potential for development. In this paper, aiming at QR code image recognition process may be encountered in noise interference, QR code accounting for the actual captured image just a part, slant, geometric distortion, and other issues have been studied, put forward through improved median filter to remove noise, use　QR code symbol characteristics to locate and rotate the tilted image, distortion which is not serious use control-point transform and spatial bilinear interpolation method for geometric correction. The results show that this method is simple and effective, can be used to improve the QR code correct recognition rate.