【作者】 李丽娜；

【导师】 杨兆选；

【作者基本信息】 天津大学 ， 信号与信息处理， 2010， 硕士

【摘要】 随着条码技术的发展,二维条码作为一种新的自动识别技术在越来越多的领域等到应用,其中DataMatrix条码以信息密度大、抗污损畸变能力强、保密性能好等优点广泛应用于多种条码易受损或某些特殊领域,如集成电路、医疗和生产自动化管理等。论文深入地研究了DataMatrix条码识读的关键技术,并在基于TMS320C6713 -DSP平台的实验系统中实现了DataMatrix条码的准确识读。论文研究的条码识读关键技术主要包括图像预处理和DataMatrix条码的解码算法,主要内容如下:⑴.在条码图像增强处理中,研究了基于小波域的图像融合方法,解决了在应用中经常遇到的因条码图像局部对比度过低而导致解码错误的问题。⑵.在条码图像二值化过程中,采用了基于小波分解和大津法的图像二值化方法,克服了整体阈值法在光照不均、噪声干扰较大的情况下处理效果较差,以及局部阈值法处理结果存在的块效应问题。⑶.采用二值形态学膨胀的方法及连通域标记的方法提取条码区域,实现了条码图像与背景图像的基本分割。⑷.在DataMatrix条码的解码过程中,采用Sobel算子对图像进行边缘检测,然后通过Hough变换获取“L”型条码边界及倾斜角度,通过网格扫描进行条码的码字识别。论文利用上述算法编写了DataMatrix条码识读程序,并在基于TMS320C6713的DSP实验系统平台上进行了实验研究。实验结果表明,论文工作中所研究的DataMatrix条码识读程序能够准确快速地识读光照不均、噪声干扰较大环境下的DataMatrix条码,并且有良好的检错纠错功能。理论分析和实验结果都表明了本文所提出的创新性算法的正确性和实际应用价值。更多还原

【Abstract】 With the development of the barcode technology, two-dimensional barcodes have seen its application as a new automatic recognition technique in more and more areas. Among the two-dimensional barcodes, DataMatrix, with high information density, strong stain-and-deformation-resistant ability and good data security as its strength, has been widely applied in multiple areas where barcodes are easily damaged or in some special domain, such as Integrated Circuit, Medical,and Automated Production Process Management and etc.This paper studied the DataMatrix recognition critical techniques, and able to decode DataMatrix used in barcode decoding experiment system which is based on the TMS320C6713 DSP platform.The paper focused on the image pre-processing process and DataMatrix decoding algorithm. The major works are as follows:⑴. In the image pre-processing part, paper researched an image fusion techniques based on wavelet domain. This method solved the problem that is often encountered in application, i.e. wrongly decoding due to extremely low local contrast.⑵. In barcode binarization, paper adoped a binarization method which is based on wavelet decomposition and Ostu method. The new method avoided both the bad performance of global threshold in unevenly illumination or strong noise conditions and the block effect in local threshold method.⑶. Adopted the dilation operation in binary morphology and the connected region labeling method to extract the barcode region, distinguishing barcodes from the background.⑷. In the DataMatrix decoding process, the paper utilized the Sobel operator to detect the edges image and then use Hough transformation to obtain the‘L’boundary and orientation of the barcode. The code words of the barcode are finally recognized through grid scanning.Experiments results show that the DataMatrix decoding techniques presented in the paper, is able to fast and accurately recognize the barcodes even in uneven illumination and strong noise conditions and, at the same time, owns high error detection and correction ability.The theoretical analysis and experiment result both demonstrated correctnes and applied value of the innovation algorithm proposed in the papar.更多还原