方便面桶包装缺陷的视觉检测技术研究

发布时间：2018-03-01 04:10

本文关键词： 方便面桶式包装缺陷检测检测软件　出处：《哈尔滨商业大学》2017年硕士论文　论文类型：学位论文

【摘要】：桶式方便面的包装过程中纸质方便面桶,会在流水生产线上经过投送、放入面饼、放入料包、压贴顶盖、压贴表面包装、塑皮封装等多道工序。在多步流水作业过程中,因为受力和摆放位置等误差,难免出现桶体发生形变、桶底出现穿孔、桶底沾染脏物等包装缺陷,这种包装质量问题如果无法有效避免,不仅将大大影响消费者的利益,也会影响到方便面生产商的经济利益。对于方便面桶式包装的缺陷检测,视觉检测系统以摄像机为核心传感器,进而以图像处理、模式识别等算法为核心软件技术,可以有效地完成方便面桶的缺陷检测。视觉检测属于典型的非接触测量,不会对方便面桶造成二次破坏;图像处理和模式识别等算法,以软件的方式完成对包装缺陷的检测,可以大大提高检测实时性和检测效率。本文开展的研究工作如下,第一,对大桶、小桶两类方便面桶进行了图像层面的描述,进而分析了其可能出现的五类缺陷,并确定本文以桶身变形、接缝缺陷、桶底破损、桶底脏污四类缺陷为主要检测对象。构建了方便面桶包装缺陷视觉检测的硬件系统,设计了方便面桶包装缺陷视觉检测的软件方案。第二,针对方便面桶包装中的桶身变形缺陷和接缝缺陷的视觉检测方法进行了研究。对检测过程中的灰度化处理、二值化处理等基本算法进行了阐述。构建了一种基于链码表格的轮廓跟踪方法。第三,针对方便面桶包装桶底的脏污缺陷和破损缺陷展开了视觉检测研究。构建了一种三阶段的预处理方法,包括均值滤波、Gauss滤波、Laplace锐化。对形态学方法进行改进,应用于带有脏污缺陷的桶底图像检测。对比了三种边缘检测方法,选取Canny边缘检测方法,应用于出现破损缺陷的桶底图像检测。第四,设计了方便面桶式包装缺陷检测软件。在检测软件界面下,对基本功能和四类缺陷检测功能进行了展示。之后,从检测时间和检测精度两个角度,对缺陷检测软件进行了性能分析。
[Abstract]:In the packaging process of barrel instant noodle, the paper instant noodle barrel will be delivered on income production line, put into flour cake, put in material bag, press top cover, press surface packaging, plastic skin package, and so on. In the process of multi-step flow, Due to the errors of force and placement, it is inevitable that the barrel body will deform, the bottom of the barrel will be perforated, the bottom of the barrel will be contaminated with dirty materials and other packaging defects. If this kind of packaging quality problem cannot be effectively avoided, it will not only greatly affect the interests of consumers. It will also affect the economic benefits of instant noodle manufacturers. For the defect detection of instant noodle barrel packaging, the visual inspection system takes the camera as the core sensor, and then takes the image processing, pattern recognition and other algorithms as the core software technology. Visual inspection is a typical non-contact measurement, and it will not cause secondary damage to the instant noodle barrel. Image processing and pattern recognition algorithms, such as image processing and pattern recognition, can be used to detect packaging defects by software. The research work in this paper is as follows: firstly, the image level description of two kinds of instant noodle bucket is given, and the five possible defects are analyzed. It is determined that the main detection objects in this paper are four kinds of defects, such as barrel body deformation, joint defects, bucket bottom breakage, and bucket bottom fouling. A hardware system for visual inspection of instant noodle barrel packaging defects is constructed. The software scheme of visual inspection of instant noodle barrel packaging defect is designed. Secondly, the visual detection method of barrel body deformation defect and joint defect in instant noodle barrel packaging is studied. The basic algorithms such as binary processing are described. A method of contour tracking based on chain code table is constructed. Third, The visual detection of dirty and damaged defects on the bottom of instant noodle barrel was studied. A three-stage pretreatment method was constructed, including the mean filtering Gauss filter Laplace sharpening. The morphological method was improved. This paper compares three edge detection methods, selects the Canny edge detection method, and applies it to the bucket bottom image detection with damaged defects. 4th, In this paper, the software for detecting the defect of instant noodle barrel packaging is designed. The basic function and the four kinds of defect detection function are demonstrated under the software interface. After that, the detection time and precision are analyzed. The performance of defect detection software is analyzed.
【学位授予单位】：哈尔滨商业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TS206;TP391.41

【参考文献】