A Correction Algorithm of QR Code on Cylindrical Surface

When capturing a QR code on a cylinder, there may be geometric distortions due to the angle of the camera and the cylindrical deformation, which makes the QR code difficult to recognize. To solve this problem, a correction algorithm for QR codes is proposed in this paper. First, the boundary of the QR code on the cylinder is extracted by a morphological algorithm and geometry calculations. Then, the feature points on the image are accurately located. Next, the standard size of the QR code is determined by using the cross-ratio method in descriptive geometry. Finally, the image is corrected by perspective projection. It is proven that the algorithm can correct the distortion of the cylinder image effectively, and it has the ability to properly identify cylindrical QR codes.