汉信码二值化和定位算法研究

发布时间：2018-11-04 09:00

【摘要】：作为我国第一个具有自主知识产权的二维码码制,汉信码的研制对提高我国条码技术应用水平、扩展二维码技术的应用领域起到至关重要的促进作用.因此,对汉信码技术的研究有着重要的现实意义.对于一幅汉信码图片,有时会出现光照不均、码图倾斜或者畸变等情况,为了将上述现象对译码的影响降到最低,需要在译码前对汉信码图片进行预处理工作.本课题在深入分析汉信码结构、编解码原则的基础上,对汉信码预处理技术中的的二值化处理以及定位做了较为细致的研究,主要工作包括以下几个方面: 首先回顾了汉信码的码制标准,编解码原则,以及RS纠错算法的理论及其在汉信码译码中的应用. 在汉信码二值化处理方面,在简单综述了全局阈值法(Otsu法,全局最优阈值迭代法等)以及局部阈值法(Bernsen法等)的基础上,给出了针对手机摄像头采集的汉信码图像进行二值化处理的算法.首先我们进行了大量的实验,基于对这些实验结果的详细分析,提出了改进的分块自适应阈值法和改进的Bernsen算法,有效克服了局部阈值法的局限性,降低了算法复杂度.相较于改进前的算法,本文在处理光照不均的汉信码图像时取得了十分明显的效果. 在汉信码定位方面,我们充分运用各种数字图像处理技术,通过对采集的汉信码图像进行腐蚀、金字塔处理、二值化、查找轮廓、凸四边形轮廓逼近等一系列处理完成对汉信码的准确定位.我们提出的算法可以成功定位出拥有多个汉信码符号的图像以及手机摄像头采集的图像.这些图像允许具有一定的倾斜角度,从而保证了定位算法的鲁棒性. 同时,我们也对增值税发票上的汉信码图像的定位进行了研究,通过计算图像旋转角度,图像校正,切割并提取码图等步骤成功实现了发票汉信码的定位,为发票汉信码的识别奠定了良好的基础.
[Abstract]:As the first QR code system with independent intellectual property rights in China, the development of Hanshin Code plays an important role in improving the application level of barcode technology and expanding the application field of QR code technology in China. Therefore, the research of Chinese letter code technology has important practical significance. In order to minimize the influence of the above phenomena on the decoding of a Chinese-coded picture, it is necessary to preprocess the picture before decoding, such as uneven illumination, skew or distortion of the code map, and so on. On the basis of deeply analyzing the structure and the principle of coding and decoding, this paper makes a detailed study on the binarization and positioning of the preprocessing technology. The main work includes the following aspects: firstly, the code system standard, encoding and decoding principles, the theory of RS error correction algorithm and its application in the decoding of Chinese letter codes are reviewed. In the aspect of binarization of Hanshin code, the global threshold method (Otsu method, global optimal threshold iterative method, etc.) and the local threshold method (Bernsen method, etc.) are briefly summarized. An algorithm for binary processing of Hanshin image collected by mobile phone camera is presented. Firstly, we do a lot of experiments. Based on the detailed analysis of these experimental results, an improved block adaptive threshold method and an improved Bernsen algorithm are proposed, which overcomes the limitations of the local threshold method and reduces the complexity of the algorithm. Compared with the improved algorithm, this paper has achieved a very obvious effect in the processing of unevenly illuminated Hanshin code images. In the aspect of Hanshin code location, we make full use of all kinds of digital image processing techniques, by corrupting, pyramid processing, binarization, searching contours, Convex quadrilateral contour approximation and a series of processing to complete the accurate location of Hanshin code. The proposed algorithm can successfully locate images with multiple symbols and images captured by mobile phone cameras. These images are allowed to have a certain tilt angle, thus ensuring the robustness of the localization algorithm. At the same time, we also study the location of the Chinese letter code image on the VAT invoice. By calculating the rotation angle of the image, correcting the image, cutting and extracting the code map, we have successfully realized the location of the invoice Chinese letter code. It lays a good foundation for the recognition of invoice code.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.44

【参考文献】