Recognition Algorithm For Large Deformation VAT Chinese-Sensible Code

Since the twenty-first century, a two-dimensional bar code has become the new darling of the information age. It is found in every corner of daily life, as a daily indispensable role for people’s food and clothing. On behalf of two-dimensional bar code system, Chinese-sensible code has a much wider application of the VAT invoice information storage, transmission and recognition. Unfortunately, limited to printers and scanners and other hardware devices of aging and wear and tear, as well as many uncontrollable factors in the image acquisition process, the Chinese-sensible code’s quality in invoice image is uneven, which seriously affecting the proper identification and information transfer. It is true that the traditional idea of the invoice recognition algorithm for Chinese-sensible code is clear and easy to implement, and for higher image quality it has reached a satisfactory invoice code pattern recognition rate, but for a large irregular deformation invoice code pattern it can do nothing. And in recent years, it is precisely the fact that invoices with such a low image quality emerge in large quantities on the market, so that people come to realize that the problem solving large deformation invoice Chinese-sensible code has become essential.Based on that situation, this paper aims at VAT invoice Chinese-sensible code, and we propose a new recognition algorithm. First, we make modular the traditional identification process and analyze the impact of the recognition rate of the problem, and retain all the reasonable modules. Next, the code pattern is divided into large deformation overall longitudinal deformation and irregular lateral deformation, both treated differently, in turn resolved. We track fixed information with gray images, and find black and white modules’ junction with the improved binary images. Finally, in accordance with the principles of the drawer, we develop one rational generating criterion, based on the image data obtained, we generate the standard code pattern corresponding to the original code pattern with a large deformation, then we can get real information directly by decoding.To demonstrate advantages of this algorithm, the experimental results will be accompanied by special. For large deformation invoices that traditional algorithm’s recognition rate can only reach 20%, the rate of this paper’s algorithm is more than 80%.