结合神经网络和投影波形的二维条码识别方法  

The two-dimensional bar code identification method combined with neural network and projection Waveform

作　　者：张菊香[1] 张薇[1] 孙腾云[1] 

机构地区：[1]西安电子科技大学,710071

出　　处：《微计算机信息》2008年第21期159-161,共3页Control & Automation

基　　金：教育部回国留学基金项目(K40604040114)

摘　　要：本文提出了一种处理矩阵式二维条码图像的新方法。首先对扫描到的图像进行除噪、滤波,然后根据投影波形旋转图像,采用折半查找法确定旋转角度,最后用Hopfield神经网络识别条码的各个矩阵点。实验证明,这种方法可以有效快速的实现二维条码图像的识别。This paper presents a new method of processing matrix two-dimensional bar code image. First we wipe off and filter the noise of the scaned images. Second rotate the images on the basis of image projection waveform, using halting search method to de- termine the rotational angle. Finally, Hopfield neural network is need to identify the various matrix bar code point, Experiments show that this method can effectively achieve rapid 2D barcode image recognition.

关 键 词：投影波形 HOPFIELD神经网络 矩阵式二维条码 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]