二维条码图像反模糊和超分辨率重建研究

As an important automatic identification technology, the two-dimensional bar code has wide range of prospects for commercial applications. Effective solution of the two-dimensional bar code image degradation is the key of its successful use. In practical applications, blurred images and low-resolution images are common degradation, and give two-dimensional bar codes recognition technology difficulty. Though existing image restoration technology has achieved good results in many fields, they aren't designed for the two-dimensional bar code and hard to solve this kind of degradation effectively. The project intends to study two-dimensional bar code image deblurring algorithm and super resolution reconstruction algorithm according to the characteristics of the two-dimensional bar code. The research mainly includes: 1) We design the regularization item using the prior knowledge of the two-dimensional bar code image, and design the deblurring algorithm based on regularization method. 2)We design the fast algorithm to optimize the object function and the adaptive method to set regularization parameter in the regularization method. 3) We combine the two-dimensional bar code image structural features and sparse representation theory to design super resolution reconstruction method based on learning. 4) We study the dictionary learning algorithm and image feature extraction algorithm in sparse representation, which is suitable for the characteristics of the two-dimensional bar code image. And we plan to apply the algorithms proposed in the project to develop a prototype system. The research results of this project can provide the theoretical basis for the development of the two-dimensional bar code recognition technology.

二维条码作为一种重要的自动识别技术，有着极其广泛的应用前景。有效地解决二维条码图像降质问题是其能够得以成功应用的关键。实际应用中，模糊图像和低分辨率采样图像是两种常见的降质，给二维条码识别带来了困难。现有的图像复原技术尽管在诸多领域取得了较好的效果，但并未针对二维条码模型设计，难以有效地解决此类图像的降质问题。本项目拟针对二维条码建模，研究二维条码图像反模糊和超分辨率重建算法。重点研究:1）应用二维条码图像先验信息构造正则项，设计基于正则化技术的反模糊算法；2）设计相应的正则化目标函数快速优化算法和正则化参数自适应设置算法；3）应用二维条码的结构特征，结合稀疏表示理论，设计基于学习的超分辨率重建算法；4）设计稀疏表示中，适合二维条码特点的字典学习算法和图像特征抽取算法。并计划将设计的算法加以应用，完成一个原型系统的设计。本项目的研究成果可为二维条码识别技术的研发提供理论基础。

二维条码是一种具有广泛商用前景的自动识别技术。有效地解决二维条码图像降质问题是其能够得以成功应用的关键。实际应用中，模糊图像和低分辨率采样图像是两种常见的降质，给二维条码识别带来了挑战。本项目重点研究二维条码图像复原技术。经过项目组的努力，已按计划并超额完成了立项规定的任务。针对模糊图像，本项目首先采用频域不变矩特征获得模糊核类型，然后结合二维条码图像特点构建了基于二值限定和稀疏先验的正则项，进步设计反模糊算法模型和基于交互迭代的求解算法。针对低分辨率的采样图像，本项目应用稀疏表示理论，根据二维条码图像结构，融合了纹理特征和方向边缘特征，设计基于学习的超分辨率重建算法。此外，为了获得高质量的学习样本，项目还研究了基于激励的学习样本采集和标注技术。.本项目发表SCI期刊论文13篇，其他期刊论文17篇，国际会议4篇；获得国家发明专利授权1项，申请国家发明专利5项，获得软件著作权2项；获得江苏省科技进步三等奖一项；完成科技成果转化投入产业，并获得国防技术应用和民用产业应用；培养硕士研究生７名，博士研究生１名。本项目的研究成果为二维条码图像复原技术的研发提供了理论基础。