基于视觉的连接器特征辨识
2020年电子技术应用第10期
乔景慧,赵校伟,何鑫达
沈阳工业大学 机械工程学院,辽宁 沈阳110870
摘要: 单目视觉识别系统中,工件的高度不同会导致测量直径尺寸的误差。通过单目相机识别图像的高度特征信息与直径特征信息,完成工件的准确测量识别。提出了一种工件多特征组合辨识的方法识别连接器。通过识别工件中心以及对工件所形成的阴影在指定的测量方向进行阴影长度测量,通过工件的高度和直径两个特征元素,识别工件的尺寸及类型。采用形态学闭运算对图像处理,而后获取工件的中心。在测量的方向上采用Halcon封装的边缘检测算法检测到亚像素边缘,而后处理获得阴影的长度尺寸。再根据组合特征确定工件类型。实验表明,所提出的工件多特征组合辨识的方法能够准确识别工件的特征,此方法能够准确区分不同工件,完成自动化生产的辨识工作。
中图分类号: TN06;TP271
文献标识码: A
DOI:10.16157/j.issn.0258-7998.191427
中文引用格式: 乔景慧,赵校伟,何鑫达. 基于视觉的连接器特征辨识[J].电子技术应用,2020,46(10):97-101,105.
英文引用格式: Qiao Jinghui,Zhao Xiaowei,He Xinda. Feature identification of connector based on vision[J]. Application of Electronic Technique,2020,46(10):97-101,105.
文献标识码: A
DOI:10.16157/j.issn.0258-7998.191427
中文引用格式: 乔景慧,赵校伟,何鑫达. 基于视觉的连接器特征辨识[J].电子技术应用,2020,46(10):97-101,105.
英文引用格式: Qiao Jinghui,Zhao Xiaowei,He Xinda. Feature identification of connector based on vision[J]. Application of Electronic Technique,2020,46(10):97-101,105.
Feature identification of connector based on vision
Qiao Jinghui,Zhao Xiaowei,He Xinda
College of Mechanical Engineering,Shenyang University of Technology,Shenyang 110870,China
Abstract: In the monocular vision recognition system, the measuring error of the diameter can be caused due to the different height of the workpiece. Accurate recognition of workpiece is completed by measurjing the height and diameter feature information of the workpiece. A methodology was proposed by integrating with multi-feature, that is the height and diameter feature.The length of the shadow in the specified measuring direction is measured by identifying the center of the workpiece.The size and type of the workpiece are identified by the height and diameter of the workpiece. The closed operation of morphology is used to process the image, and then the center of the workpiece is obtained. In the direction of measurement, the edge detection algorithm with Halcon package is used to detect the subpixel edge, and then the length of the shadow is obtained by post-processing. The workpiece type is determined according to combining features. A test show, the method of multi-feature combination identification of the workpiece proposed in this paper can accurately identify the characteristics of the workpiece, and this method can accurately distinguish different workpiece to complete the identification of automatic production.
Key words : monocular vision;connector;workpiece recognition;morphological closed operation;Halcon
0 引言
目前在产品的自动化生产中视觉系统占据了很大的比重。采用图像的测量技术不需要直接接触测量对象效率高,同时视觉系统的采用可以极大提高自动化程度[1-2],图像的采集、处理、分析可由计算机程序设定,极大地提高生产效率[3-4]。而在自动化生产领域,视觉系统应用范围很广,应用范围涵盖到工业元件的尺寸测量、物体形状检测、模式识别、模板匹配等诸多方面[5-6]。但是目前自动化生产中所采用的视觉系统,由于存在相机和镜头加工误差、装配误差、光源选用,都会对检测结果造成影响,影响到视觉系统的检测精度[7]。
在多产品的自动化生产中,对工件的辨识是至关重要的步骤[8],本文针对自动化生产的连接器组件的类型辨识进行研究。在不同型号的连接器类型中,有一些型号直径尺寸特征相近,在对其采取直径特征检测时,不同测量时多方面因素影响而产生的误差,导致对直径的测量出错。而直径尺寸相近的连接器,高度有明显的区别,尺寸相差较多。
本文针对上述情况,提出通过连接器的组合特征辨识连接器的类型,采用单目相机通过对工件进行单侧光源投影,利用产生的阴影,结合几何数学的思想,对工件的尺寸和种类进行识别,其中通过形态学识别工件中心,其次确定测量矢量方向,沿着矢量方向通过边缘检测得出工件尺寸测量结果,辨识出工件种类。
本文详细内容请下载:http://www.chinaaet.com/resource/share/2000003031
作者信息:
乔景慧,赵校伟,何鑫达
(沈阳工业大学 机械工程学院,辽宁 沈阳110870)
此内容为AET网站原创,未经授权禁止转载。