综合利用投影算法和相似距离算法的二维条码识别

TWO-DIMENSIONAL BARCODE RECOGNITION USING BOTH PROJECTION ALGORITHM AND SIMILARITY DISTANCE ALGORITHM

文中以PDF417为例研究了二维条码的识别算法。根据Hough变换定位条码图像后,利用双线性插值将条码旋至水平,然后在条码中定位单个符号字符的图像,并充分利用图像二维特性,提出投影算法识别单个符号字符。最后根据识别出来的码字讨论了条码符号字符的相似距离识别算法。文中根据提出的距离函数来确定正确的码字,并分析了相似距离法对首读率和正确率的影响。实验结果表明综合利用投影算法和相似距离算法具有良好的性能,显著地提高了条码的识别率,满足了实际使用的要求。

Taking PDF417 as an example, the recognition research of two-dimensional barcode is discussed in this paper. Using Hough Transform, the position of the barcode is located in the image and the barcode is rotated to be horizontal based on bilinear interpolation. Then the image is further segmented to get the single codeword image and it is recognized by projection algorithm. At last, the similarity distance recognition algorithm of codeword is discussed as an important topic in the paper. The correct codeword is decided based on similarity distance function proposed in the paper. First-reading rate and correct rate of the similarity distance algorithm is also analyzed here. The result shows that using projection algorithm and similarity distance algorithm can get good performance. It can achieve high recognition rate, which suffices to the requirement of the practical use.