Research And Realization Of The Recognition Algorithms Of Two-Dimensional Character Barcode Based On Scanner

With the rapid development of barcode technology and advent of the era of mobile information, the application of barcode technology become more and more widespread, and the mobile electronic tickets that combine barcode with mobile terminal are going to be the trend. A new barcode called2D Character Barcode in recent years rised, which is different from traditional graphic2D barcode, consisted of English letters. And this barcode can be sent via SMS, which not only reduces costs but also improves the success rate of the receiver, will be an important part of mobile value added services.This paper analyzes the structural characteristics of2D Character Barcode, and analyzes the system architecture and application development of scanner terminal, based on that and improved barcode location algorithm, characters segmentation and characters recognition to complete and implement scanner2D Character Barcode recognition system based on scanner. The main contents are as follows:1. On the location of barcode region, with regards to the geometric features of the locator of2D character barcode image and memory restrictions of scanner, an improved circularity detecting method is designed based cross algorithm, Shape Angle value algorithm and circularity calculation algorithm to extract the location of barcode area. It will not only make the location more quickly, but also more accurate.2. On the character segmentation, based on the design and arrangement of the character barcode rules, proposed a character segmentation method based on projection techniques. Firstly, smoothed the histogram twice to split roughly, and then subdividing based on the pixel statistical methods. This method takes into account the overall distribution of the character and the local distribution of features, makes segmentation more accurately.3. On the character recognition, selected16easilydistinguishable characters as barcode elements. These characters are selected bymaximum between class distance method and minimum correlation coefficient method. According to the local, global and structural characteristics of the characters to extract the characters, and ultimately used the characteristics template matching method for recognition.The algorithms generate a specific plug-in format by compiler, and loaded to the Honeywell’s1900GSR scanner. Then test it by scanning the actual2D Character Barcode, achieving real-time decoding well.