【作者】 李鹏；

【导师】 朱宏辉；

【作者基本信息】 武汉理工大学 ， 机械电子工程， 2004， 硕士

【摘要】 随着国民经济的飞速发展和经济全球化的到来，对物流仓储系统的需求越来越大，而物流仓储系统的信息化和自动化对减少企业运营成本，提高企业生产效率起到了巨大的推动作用。我们知道，物流信息自动识别技术是物流自动化系统可靠运行的前提，它是集光、电、计算机等技术为一体的高新技术，是数据自动采集、自动输入的基础，是计算机“实时”处理的重要技术保障，在现代物流系统中发挥着极其重要的作用。因此，对物流系统中信息识别技术的研究有着重要的意义。 图像识别与处理长期以来都是备受关注的研究课题，其研究成果甚多，应用领域十分广泛，作者在分析了当前图像处理技术在模式识别中的发展和信息识别技术在物流仓储中的应用后，结合其优缺点，提出了利用对象的图像信息进行物流模式识别这一研究内容。 根据人类认知的逻辑模式，论文的整个体系主要从两个部分加以探讨： 第一部分是外形的识别。该部分主要论述了图像不变矩的原理和应用现状，分析了在离散状态下比例因子对7个不变矩的影响，提出了一种的改进算法，并利用其提取了物体外形轮廓的特征向量，最后用BP网络进行了模式分类。当中涉及的不变矩特征量可以用来表征图像目标的外形特征，联系实际应用，我们可以用该方法来检测货物包装外形的损坏情况。对现有物流识别技术而言，该方法可以作为一个有益的补充。 第二部分是字符的识别。主要论述了字符识别技术的方法、原理和研究现状。此外，作者结合当前字符识别中的一些常用方法，使用数学分析软件MATLAB进行了具体实现。字符中包含着丰富的信息量，因此利用字符识别方法可以提取货物包装上的文字信息对物体进行分类。 识别技术的应用与实际场景密切相关，由于作者受试验条件和本身精力的限制，数据的采集有一定的局限性，研究的结果要应用到实际中还需要改进。因此，论文最后在总结了前期的工作的基础上，对进一步的工作提出了建议和展望。更多还原

【Abstract】 With the rapid development of national economy and the emergence of the economic globalization, logistics storage system is in increasing demand. And its informationization and automatization play an important role in reducing enterprises’operation cost and improving enterprises’ productivity. We know that logistics information automatic recognition technology is the prerequisite that automation system run reliably. It is the new and high technology that incorporates many different disciplines, such as optics, electricity, computer and so on. It is the foundation that gathers and inputs data automatically. It is the important technological guarantee of computer real-time work. Now, it is playing more and more important role in modern logistics system. Therefore, the study of information recognition technology in logistics system is of great importance.Image processing and recognition are received much concern for a long time. Lots of research results are acquired in this field and have been applied in many aspects After analyzing the development of current image processing technology in the field of pattern recognition and the application of information recognition technology in logistic storage, the author puts forward a point that the image information of a target should be utilized for logistics pattern-recognitionAccording to the logic mode of humankind’s cognition, the thesis consists of two main parts:The first part is about the appearance recognition. In this part, the author mainly introduces the theory of the image moment invariants and its current application, analyses the influence of scale factor on moment invariants features under the state of discrete and adopts a sort of modified algorithm to abstract the eigenvector of the objects’ outline Finally, the author uses BP network to classify the mode. Since the moment invariants can signify the appearance feature, it can be used to detect the goods’ damage status in practical application As far as the existing logistic recognition technology is concerned, this methodis very helpful.The second part deals with character recognition The method, principle and current research status of character recognition are discussed. In addition, combined with some commonly used methods, MATLAB-a mathematics analytical software-is used to realize the character recognition. Characters contain abundant information, so we can draw the character information on the package of goods to classify the objects.The application of recognition technology is closely related to practical conditions. Because the author is restricted by experimental condition and his own energy, the collection of data is restricted to some extent. And there is room for improvement of the research result before it is applied to practice. Finally, some suggestions and expectation for further study are put forwarded based on a summary of the research at earlier stage.更多还原