Study On Recognition Of Two Dimensional Bar Code Of Key Technologies

Bar code technology which was born 20 years ago has been developing rapidly recently. It has been applied widely in many fields, it is economic, practical automatic recognition technology Company with the development of information technology and globalization, 2-D bar code came out with the characteristics of high density, storage and dependability. The most popular 2-D bar code is PDF417.Now, almost all of recognition devices of 2-D bar are expensive because they are imported from oversea. But those devices can't process the image without key information and the image with high noise well. Although more and more experts take part in this research field, the result is not satisfaction.The image of 2-D bar code is classified into two types: one is the image without key information (start pattern symbol, end pattern symbol or row and column indication symbol) and the other is the image with full of key information. In this paper, we propose a practical and feasible recognition algorithm to process PDF417 image with high noise by using image process and pattern recognition technology and it includes two methods according to the types of 2-D image. The main differences between the two methods are how to locate the area of bar code, how to obtain the number of the row and column and whether it can recognize the rotated images. The procession of the first method is: after effective bin-process, 2-D code bar area can be located by searching the start pattern and end pattern from the horizontal and vertical directions, and then column lines are found in this area. We also can obtain the value of row, column and error level by line scan in the row and column indication area which need geometry deforming. In addition, we use these values to divide bar code area into code-word cell images which are also need deforming, finally, we use the special analyzing method to get the code-words and get the data information by explaining. The second method makes full use of the characteristics of PDF417 symbol to obtain number of the row and column because the image lose the key information, the rest steps are just the same as the first method.The experiment proves our algorithm is feasible and effective.