Recognition Algorithm And Implementation For PDF417 Code On The CD

With the rapid development of Internet of things, the study of information storage and information transmission is more and more popular. The bar code technology because of its simple and efficient way has been widely used in various fields of social production and social life. Due to capacity limits, traditional one-dimensional bar code cannot meet the application requirements. Two dimensional barcode arises at the historic moment represented by QR code, Chinese-sensible code and PDF417. In particular, with large information capacity, wide range of information coding, decoding high reliability advantages, PDF417 has already aroused people’s high attention. On the current market, there has already some PDF417 bar code recognition technology which is monopolized by foreign technology or can only process under ideal conditions of bar code identification. According to the wide use of PDF417 on CDs, this paper researches on the CD’s PDF417 code identification with complex background or barcode serious tilt condition and it turns out that we obtained the corresponding results. In this paper, our study is mainly divided into the image preprocessing,barcode recognition and decoding three parts.The image preprocessing part is divided into background processing, image skew correction and code figure extraction three stages. During the specific application process, in order to get rid of the background interference problems, we first lock range to narrow the barcode area roughly which reduces amount of calculation for subsequent processing. Then find the interference area of background using connectivity and replace it by normal background. Next use gray integration to narrow the effective barcode area again, which might be used in the following image processing when necessary to reduce amount of calculation as well. For image skew problem, we first use Hough transform to detect the barcode edge. Due to the discontinuity of code figure, do open operation and expansion and then seek the slant angle of edge line. But the large computation of Hough transform bring huge burden to algorithm so that the precision requirements of slant angle is low. Next use gray integration to find the accuracy of slant angle. Combining the two methods, we can correct image to any given precision. In the stage of code figure extraction, we also use gray integration to extract the target barcode.Image recognition process is divided into modules grids part and module gray attribute judgment part. Mainly combining with the characteristics of PDF417 code,the former make use of gray integral statistics and calculation module width to determine module boundaries and obtain grid lines. About the module gray attribute judgment part, due to the advantages of the experimental images’ large amount of information as well as small errors, this paper only uses the module’s average gray to determine the threshold. In this way, judgment module gray-scale properties can be achieved.Decoding part can be done just according to the rules of coding and corresponding algebraic operation.Experimental results show that the recognition algorithm of CD’s PDF417 code has feasibility and effectiveness.