QQ客服 官方微博 反馈留言

题名	食品密封包装的微孔识别与差异性分析
作者	张倩
学位类别	硕士
答辩日期	2014-05-28
授予单位	中国科学院沈阳自动化研究所
导师	马钺
关键词	小波包 经验模态分解 总体平均经验模态分解
其他题名	Research on pinhole recognition and variance analysis for food sealing package
学位专业	模式识别与智能系统
中文摘要	在当前的日常生活中，食品安全问题越来越受到国民的关注。在密封食品的包装过程中，由于漏封、压穿或材料本身存在的裂缝的问题总会形成内外连通的小孔，这都会对包装内容物产生不利影响，是威胁食品安全的主要问题之一。一旦食品包装的密封性不好，食品的质量将失去保证，这不仅将对企业造成一定的经济损失，更对人们的身体健康造成威胁。目前，密封试验主要用在电子行业，而在软包装行业，密封监测主要通过手工监测和浸水的方法实现，这种方法工作效率不高，耗时，而且对于很微小的小孔很容易发生漏检的现象。 在食品和医药行业，检测塑料容器的密封性能可以采用高压放电技术，这种技术可以找到容器的各种渗透、漏封问题，是当前使用的非接触式检漏方法中的一种新型高可靠性方法。目前，国内外对介质阻挡放电机理、应用和放电电源都做了大量研究。由于密封食品在不同破损状态下的阻抗性存在差异，当高频高压电经过密封食品时，不同破损状态的密封食品的负载电压不同。根据这一原理，以火腿肠为例，本文提出了一种新的基于介质阻挡放电的检测方法，设计了一个新的检测思路，构建了一个基于支持向量机和总体平均经验模态分解的火腿肠放电信号针孔差异性识别系统。 首先，根据放电信号的非平稳特性以及小波包变换的多分辨率分析在信号去噪中的独特优势，介绍了小波和小波包变换的基本原理，仿真研究在小波包阈值去噪中不同参数选择对信号去噪效果的影响，以此为参考进行参数选择，实现对放电信号的去噪处理。然后，从密封食品介质阻挡放电的基本原理出发，引入总体平均经验模态分解（EEMD）方法对放电信号进行分解，将各IMF分量的能量比和能量熵作为抽取的特征值，仿真验证了其重复性和差异性，并与经验模态分解（EMD）提取效果进行对比，证明了总体平均经验模态分解的抗混叠能力。最后介绍了支持向量机的分类原理，将能量比和能量熵作为支持向量机的输入量进行放电信号的状态识别，并与BP神经网络分类结果进行对比，证明了支持向量机在分类精度、训练时间上的优越性。 本文将EEMD方法与支持向量机相结合的方法运用到密封食品的密封性检测中，实现了不同状态放电信号的分类，验证了这种检测方法的可行性，为后续的这种检测方法的实际应用提供了理论依据。
索取号	TP391.4/Z33/2014
英文摘要	The issue of food safety has already became the focus of public attention. In the process of packaging, due to the influence of various factors such as leakage sealing and pressure to wear, forming connecting holes inside and outside, these are harmful to the packing contents . As a result, the damaged packaging is one of the major problems to threaten sealing food security. Once the sealing of vacuum packaging and inflatable packaging is not good, the quality of the food will lose the protection, which not only will cause certain economic losses to the enterprise, but also will be a threat to people's health. At present, the sealing test is mainly used in electronic industry. In the flexible packaging industry, sealing monitoring is realized mainly through manual monitoring and water soaking, and these methods are inefficient, time-consuming and prone to failure phenomenon for the pinholes. In food and pharmaceutical industries, the use of high voltage discharge to detect the sealing performance of plastic container, can find the microleakage in plastic container, which is a new non-contacting and reliable detection method. At present, the domestic and foreign experts have done a lot of research in terms of the principle and application of dielectric barrier discharge and the discharge power supply. Due to the impedance difference of sealing packaging food in various states of disrepair, when high voltage is across sealing food, the food in different damage states has different load voltage. According to this principle, ham sausage, for example, this paper proposes a new detection method based on the dielectric barrier discharge, designs a new test idea, and also builds a recognition system of discharge signals differences based on SVM and EEMD. Firstly, according to the non-stationary characteristics of discharge signals and the unique advantage of wavelet packet transform in signal de-noising, it analysis the principle of wavelet and wavelet packet transform, and through the simulation it is verified that different parameter selections in wavelet packet threshold de-noising have an impact on the signal de-noising effect. As a reference for parameter selection, the discharge signal de-noising is realized. Secondly, based on the principle of dielectric barrier discharge for sealing food, EEMD method is introduced to achieve the discharge signal decomposition, and energy ratio and energy entropy of each IMF component is extracted as the characteristic value. Their repeatability and diversity are verified by simulation. Comparing with the extracting effects of empirical mode decomposition (EMD), it is proved that EEMD has the ability of eliminating modal aliasing. Finally, it introduces the theoretical characteristics of support vector machine (SVM). The energy ratio and energy entropy used as input of support vector machine (SVM) is to realize the recognition of the discharge signal state. Through comparing of BP with support vector machine classifier in discharge signal identification it is found that support vector machine shows good superiority in recognition accuracy, training time and reliance on the model structure． This paper combines EEMD decomposition method and support vector machine for sealing detection of sealing food，and achieves the classification of discharge signals under different conditions. And it also verifies the feasibility of this test method, and provides a theoretical basis for the practical application of this detection method.
语种	中文
产权排序	1
页码	63页
分类号	TP391.4
内容类型	学位论文
源URL	[http://ir.sia.ac.cn/handle/173321/14779]
专题	沈阳自动化研究所_智能检测与装备研究室
推荐引用方式 GB/T 7714	张倩. 食品密封包装的微孔识别与差异性分析[D]. 中国科学院沈阳自动化研究所. 2014.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。