大批量小零件自动给料和图像检测系统设计与研究
发布时间：2018-09-07 10:47
【摘要】：零件在大批量、连续自动化生产时，由于各种因素的影响，总会有残次品的存在。传统的人工抽样检测方法，无法保证零件百分百的合格。随着计算机技术的快速发展，利用图像处理技术对零件进行自动检测的方法近年来得到广泛的应用和发展。 为了解决零件人工检测效率低、精度低的问题，本文设计开发了一个运用机器视觉技术对大批量小零件自动进行检测、剔除不合格零件的系统。系统的主要功能包括自动给料、图像处理检测、零件传送、废品剔除、系统控制和人机交互。 本文主要对系统的自动给料定向功能、系统控制及人机交互功能和图像处理检测功能进行了研究。在对自动给料定向功能的研究中，分析了零件在振动盘螺旋轨道上的受力情况，描述了零件在振动盘中的进给机理，并设计了定向系统，使零件可以按照需要的姿态传送给图像拍摄模块。在系统控制和人机交互的研究中，本系统采用西门子PLC作为系统逻辑控制器，针对系统的逻辑控制需要，设计了PLC外部接线和内部运行程序，为了满足系统的人机交互的功能，运用了触摸屏对系统进行组态，监控系统的工作过程。在图像处理检测功能的研究中，本文通过分析了缺陷零件的特点，提出了检测方法，即首先采用双层链表法实现对零件图像连通区域的标记，然后采用极半径不变矩对零件图像进行形状描述，再与合格零件图像的每个特征对图像中心的标准化欧氏距离以及阈值进行比较，最后通过判断被测零件图像的特征个数以及每个特征的标准化欧氏距离，从而达到检测零件的目的，检测软件由LabView实现。最后，本文运用Pro/E软件对系统进行了三维建模，并实现了整个系统的机构动态仿真。 本系统运用PLC可编程逻辑控制器作为系统控制器，机器视觉和图像处理作为检测方法，运用触摸屏和Siemens WinCC组态软件作为监控系统，可以提高整条生产线的自动化程度、降低劳动强度，提高零件检测的产品质量和生产效率。由于系统采用一种非接触的检测方式，可以更好地保证最终产品品质。
[Abstract]:Parts in mass, continuous automatic production, due to the influence of various factors, there will always be defective products. Traditional manual sampling testing method can not guarantee 100% qualified parts. With the rapid development of computer technology, the method of automatic detection of parts using image processing technology has been widely used and developed in recent years. In order to solve the problems of low efficiency and low precision of manual inspection of parts, this paper designs and develops a system which uses machine vision technology to automatically detect large quantities of small parts and eliminate unqualified parts. The main functions of the system include automatic feeding, image processing and detection, part transfer, scrap removal, system control and man-machine interaction. In this paper, the functions of automatic feeding orientation, system control, man-machine interaction and image processing are studied. In the study of the function of automatic feeding orientation, the stress of the parts on the spiral track of the vibration disk is analyzed, the feeding mechanism of the parts in the vibration disk is described, and the orientation system is designed. The parts can be transmitted to the image shooting module according to the desired posture. In the research of system control and man-machine interaction, the system adopts Siemens PLC as the logic controller of the system. According to the need of the logic control of the system, the external wiring of PLC and the internal running program are designed to meet the function of man-machine interaction of the system. The touch screen is used to configure the system and monitor the working process of the system. In the research of image processing and detection function, this paper analyzes the characteristics of defective parts, and puts forward a detection method, that is, the double-layer linked list method is used to mark the connected area of the image of parts. Then the shape of the part image is described by pole radius invariant moment, and then the standardized Euclidean distance and threshold value of the image center are compared with each feature of the qualified part image. Finally, by judging the number of features and the standardized Euclidean distance of each feature, the detection software is implemented by LabView. Finally, this paper uses Pro/E software to model the system, and realizes the dynamic simulation of the whole system. The system uses PLC programmable logic controller as system controller, machine vision and image processing as testing method, touch screen and Siemens WinCC configuration software as monitoring system, which can improve the automation of the whole production line. Reduce labor intensity, improve product quality and production efficiency of parts inspection. Because the system adopts a non-contact detection method, it can better guarantee the final product quality.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP274;TH237.1
本文编号：2228045
[Abstract]:Parts in mass, continuous automatic production, due to the influence of various factors, there will always be defective products. Traditional manual sampling testing method can not guarantee 100% qualified parts. With the rapid development of computer technology, the method of automatic detection of parts using image processing technology has been widely used and developed in recent years. In order to solve the problems of low efficiency and low precision of manual inspection of parts, this paper designs and develops a system which uses machine vision technology to automatically detect large quantities of small parts and eliminate unqualified parts. The main functions of the system include automatic feeding, image processing and detection, part transfer, scrap removal, system control and man-machine interaction. In this paper, the functions of automatic feeding orientation, system control, man-machine interaction and image processing are studied. In the study of the function of automatic feeding orientation, the stress of the parts on the spiral track of the vibration disk is analyzed, the feeding mechanism of the parts in the vibration disk is described, and the orientation system is designed. The parts can be transmitted to the image shooting module according to the desired posture. In the research of system control and man-machine interaction, the system adopts Siemens PLC as the logic controller of the system. According to the need of the logic control of the system, the external wiring of PLC and the internal running program are designed to meet the function of man-machine interaction of the system. The touch screen is used to configure the system and monitor the working process of the system. In the research of image processing and detection function, this paper analyzes the characteristics of defective parts, and puts forward a detection method, that is, the double-layer linked list method is used to mark the connected area of the image of parts. Then the shape of the part image is described by pole radius invariant moment, and then the standardized Euclidean distance and threshold value of the image center are compared with each feature of the qualified part image. Finally, by judging the number of features and the standardized Euclidean distance of each feature, the detection software is implemented by LabView. Finally, this paper uses Pro/E software to model the system, and realizes the dynamic simulation of the whole system. The system uses PLC programmable logic controller as system controller, machine vision and image processing as testing method, touch screen and Siemens WinCC configuration software as monitoring system, which can improve the automation of the whole production line. Reduce labor intensity, improve product quality and production efficiency of parts inspection. Because the system adopts a non-contact detection method, it can better guarantee the final product quality.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP274;TH237.1
【参考文献】
相关期刊论文 前10条
1 张剑剑;朱荻;李寒松;;基于数字图像处理技术的微小群孔快速检测系统[J];传感器与微系统;2009年06期
2 沙毅,曹英禹,王经武,盛国松;基于图像处理技术的工件长度在线测量[J];东北大学学报;2005年10期
3 常春国;徐运涛;;应用机器视觉精确检测工件尺寸的研究[J];电子质量;2008年02期
4 林长青;庄石;廖俊必;;基于机器视觉的螺钉在线检测[J];工具技术;2008年04期
5 曹茂永,孙农亮,郁道银;用于模式识别的极半径不变矩[J];计算机学报;2004年06期
6 任靖,李春平;最小距离分类器的改进算法——加权最小距离分类器[J];计算机应用;2005年05期
7 沈乔楠;安雪晖;;基于游程递归的连通区域标记算法[J];计算机应用;2010年06期
8 林创荣;高健;陈新;;基于网格模型的圆度在线检测系统开发[J];机械设计与制造;2011年04期
9 范彦斌,李西兵,沈自林;基于亚像素的边界检测与提取算法及误差分析[J];机械科学与技术;2005年02期
10 杨树林;陈雁秋;;基于图像表征特性的轮廓线提取的研究[J];计算机应用与软件;2008年07期
相关硕士学位论文 前3条
1 张学友;二自由度并联机器人开发及虚拟示教控制技术研究[D];南京航空航天大学;2011年
2 王海霞;基于不变矩的目标识别算法研究[D];中国科学院研究生院（长春光学精密机械与物理研究所）;2004年
3 张留刚;基于机器视觉技术的烟条检测系统研究[D];南京航空航天大学;2008年
，本文编号：2228045
本文链接：https://www.wllwen.com/kejilunwen/jixiegongcheng/2228045.html
