基于机器视觉的汽车碳刷表面检测系统设计

发布时间：2018-06-17 10:36

本文选题：机器视觉 + 碳刷表面　；参考：《江苏大学》2016年硕士论文

【摘要】：本课题所研究的对象是汽车上面用到的各种电机碳刷,碳刷的性能优劣将直接影响到各种电机的运转稳定性和效率,正是因为碳刷是绝大部分直流电机不可或缺的关键部件,对碳刷生产过程中的在线检测准确性和高效率必须要得到有力的保证。在各种工业场合特别是汽车微电机上,碳刷是必不可少的重要零部件,使用特别广泛。因此,一旦碳刷存在表面缺陷问题则会带来轻则停机重则伤及人身安全的重大事故。为确保有效降低安全事故的发生,必须在碳刷生产过程中严格把控碳刷质量,做到准确检测并剔除不良碳刷。经对比分析国内外碳刷表面缺陷检测方法的优劣势,最后给出一套实时、非接触高效准确检测碳刷表面质量的解决方案。该解决方案是基于智能机器视觉的快速发展而设计的,其优点是非接触、快速、准确,同时在应用的电子环境中具有较强的抗干扰能力,所以在各种工业生产中应用越来越广泛。为实现机器视觉准确快速检测碳刷表面缺陷的效果,本文主要做了以下几方面的工作:首先通过对比人工肉眼检测、传统在线检测、智能视觉系统非接触检测各自特点,分析了人工肉眼检测和传统在线检测的缺点和不足,同时提出了采用智能视觉系统配合改进硬件装置实现非接触在线100%高效检测碳刷的全自动方案。根据碳刷表面裂痕检测系统的指标要求,主要包括精度要求、错检率和漏检率要求以及效率要求等,对整个检测系统的硬件部分进行全面设计和选型,基于系统运行稳定性和可靠性考量来对每个环节零部件进行合理分析与选择,针对视觉系统拍照过程中存在的相机微动情况对相机固定支架架设方式进行了改进;针对振动理料送料装置存在卡料且送料慢问题增加了直线送料器,从而保证自动送料的稳定和快速;在机械翻转机构配合下实现碳刷正反面的裂痕检测;根据碳刷表面裂痕的特点和客户对自动检测装置的性能要求,对智能机器视觉系统中涉及到的几个关键部件如重点对智能相机、光源控制器、镜头、光源架设方式等进行详细分析与对比选择,最终将视觉系统中每个关键部件合理地确定下来。基于以上硬件系统和光学系统的选型确定后,对整个碳刷表面裂痕缺陷检测自动线的控制系统进行详细设计,同时对该系统的控制器以及外设电气部分也予以对比择优选择。为保障整个碳刷表面裂痕检测装置的运行稳定,对检测系统的保护装置进行设计和优化,同时设计了碳刷良品和不良品自动分拣机构;本设计的创新之处除了首次在碳刷行业引入智能视觉系统进行非接触100%检测,还引入监控系统对整套自动检测装置进行实时监控。最后,根据碳刷表面裂痕缺陷的特点,设计了采用快速检测碳刷表面裂痕缺陷的算法和处理软件。本碳刷表面实时检测方案已付诸实施,设计安装调试完成了自动在线检测装置,给一家碳刷行业龙头企业的生产线上使用本检测系统已取得了良好的效果,可以实时准确判定出碳刷表面有肉眼能看得到的所有裂痕,装置运行时检测速率有50只/分钟,比之前耗人耗时且品质无法保证的情况有了根本性改观。目前国内碳刷企业采用机器视觉在线检测表面缺陷的尚比较少,因此本装置具有较广泛的应用价值,同时本装置通过更改检测算法后亦可推广应用于其他精密五金产品的在线检测中。
[Abstract]:The object of this study is the various motor carbon brushes used in the automobile. The performance of carbon brush will directly affect the stability and efficiency of all kinds of motor. It is precisely because carbon brush is the key component of most DC motor. The accuracy and efficiency of on-line detection in the process of carbon brush production must be obtained. Strong assurance. Carbon brushes are essential components in a variety of industrial applications, especially automotive micromotors, and are used especially widely. Therefore, once the surface defects of carbon brushes exist, there will be significant accidents with light weight stopping and physical safety. To ensure the effective reduction of safety accidents, carbon brushes must be produced. In the process, the quality of the controlled carbon brush is strictly detected and eliminated. The advantages and disadvantages of the surface defect detection methods at home and abroad are compared and analyzed. Finally, a set of real-time, non-contact and efficient solution for the accurate detection of the surface quality of carbon brush is given. The solution is designed based on the rapid development of intelligent machine vision. The advantages are non-contact, fast, accurate, and have strong anti-interference ability in the applied electronic environment, so it is more and more widely used in all kinds of industrial production. In order to realize the accuracy and rapid detection of the surface defects of the brush surface, the following work is done in this paper: first by comparing artificial eye detection, transmission. The defects and shortcomings of artificial naked eye detection and traditional online detection are analyzed by on-line detection and non contact detection of intelligent visual system. A full automatic scheme using intelligent visual system and improved hardware device to realize non-contact on-line 100% high efficiency detection of carbon brush is proposed. According to the index of the carbon brush surface crack detection system Requirements, mainly including precision requirements, error detection rate and leakage rate requirements as well as efficiency requirements, all the hardware parts of the whole detection system are fully designed and selected. Based on the stability and reliability of the system, the reasonable analysis and selection of the parts of each link are carried out, and the camera micro motion in the process of visual system photography is made. The method of fixing the fixed frame of the camera is improved, and the linear feeder is added to the material feeding device of the vibrating material feeding device and the feeding is slow, thus ensuring the stability and speed of the automatic feeding, and the crack detection of the carbon brush positive and negative surfaces is realized under the coordination of the mechanical turnover mechanism; according to the characteristics of the surface cracks of the carbon brush and the customer's self The performance requirements of the dynamic testing device are analyzed and contrasted with the key components involved in the intelligent machine vision system, such as the intelligent camera, the light source controller, the lens, the light source erection mode and so on. Finally, each key component in the visual system is reasonably determined. Based on the above hardware and optical systems, After the selection is determined, the control system of the surface flaw detection automatic line of the carbon brush surface is designed in detail, and the controller and the peripheral electrical part of the system are also selected. In order to ensure the stability of the whole carbon brush surface crack detection device, the protection device of the detection system is designed and optimized. In addition to the design of carbon brush and defective products automatic sorting mechanism, the innovation of this design in addition to the first in the carbon brush industry, the introduction of intelligent visual system for non-contact 100% detection, and the introduction of monitoring system for the complete set of automatic monitoring device for real-time monitoring. Finally, according to the characteristics of the carbon brush surface flaw defects, the design of rapid detection is adopted. The real time detection scheme of the carbon brush surface has been put into practice. The automatic on-line detection device has been completed by the design and installation and debugging. It has achieved good results for the use of this detection system on the production line of a carbon brush industry leading enterprise. It can accurately determine the surface of the brush on the carbon brush in real time. In all the cracks, the detection rate of the device is 50 / minute, compared with the time consuming and the quality can not be guaranteed. At present, the domestic carbon brush enterprises are still relatively small in the use of machine vision to detect surface defects on line. Therefore, this device has a wide application value, and the device has been checked by change inspection. It can also be applied to other precision metal products on-line testing.
【学位授予单位】：江苏大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：U468.2;TP391.41

【参考文献】