The two-dimensional code image tilt correction method based on least squares support vector machines

In order to solve the two-dimensional code image tilt problem which affects locating the image, establishing and identifying sampling network, we suggest a two-dimensional image tilt correction method. In the proposed method, the least squares support vector machine (SVM) is used to regress the pixel coordinates on the two-dimensional code contour line, which can calculate the two-dimensional code image tilt vector and offset angle. This method converts the searching process for image tilt angle to solving linear matrix equation directly, which simplifies calculation and improves the efficiency of the algorithm. Also it avoids randomness and uncertainty in the process of searching. The simulation results show that this method can achieve good results for images that contain noise interference.