Abstract
QR codes are the most widely used in two-dimensional barcodes nowadays. Image pre-processing is an important process to achieve QR code image recognition in complex conditions and binarization that directly determines the success or failure of the automatic image recognition technology is a key step in the image pre-processing method. In practical applications, uneven illumination takes a heavy toll on the QR code image captured by video cameras. This paper presents an improved QR code image binarization method to deal with the QR code image with uneven illumination. The method is based on an adaptive local binarization algorithm proposed by J. Sauvola. In addition, the improved method adjusts the selected threshold derived from the improved Sauvola's algorithm and utilizes an integral image to shorten the calculation time. Experimental results show that the QR code binarization can be done quickly by the improved algorithm with high accuracy and stability, and it also can be implemented in embedded systems.