The Study Of 2D Barcode Recognition Based On Image Processing

The two-dimensional bar code is a new automatic identification technology which created along with the development of information technology. QR code (quick response code), a matrix two-dimensional bar code, is widely used in our country, according to its efficient encoding characteristics of Chinese characters and rapid identification. The information of two-dimensional bar code is high density, and the recognition technology is complicated. So, it is of great theoretical and practical significance to study the problems of barcode recognition which get form CMOS camera equipment.This thesis mostly studies the recognition of QR code based on image procession. It improved a variety of methods for identifying QR Code, and realized extracting QR code from background accurately. The main ideas are as follows:Firstly, introduces the QR Code symbol structure, encoding and decoding methods, provides a theoretical basis of the fellow two chapters.Secondly, For the QR code is collected by ordinary camera equipment, the background is uneven illuminated, an adaptively method is used to compensate the illumination. Then we use OTSU for image binarization. The problem of QR Code image recognition under different illumination condition can be solved well.Thirdly, QR code may be rotated and distorted, in this paper the corner points been found by utilizing QR code's finder pattern pixel rules, Inverse perspective transform and an appropriate interpolation method are used to correct image. At last, we generate the cell grids self-adaptively, and QR Code is conversed into a matrix composed with"0"and"1".Lastly, when the position detection pattern contamination happened, the general method is invalidly. This paper set up a new module, which composed with three finder pattern. The team progress algorithm is used for identify the best matching location, and finally finds out areas where the QR code is. Tests show that the method can expand the use of QR code.