视觉检测系统检测高亮反光零件隐蔽缺陷的方法
发布时间:
2023-02-10 11:30
我们都知道视觉检测系统在缺陷检测方面有着突出的贡献,但是对于那些高光、反光的零部件缺陷检测是否具有同样的效果呢?根据红宝多年的视觉检测经验发现:在检查光亮、反光的零件是否有缺陷,需要采用新颖的照明和成像技术的自动成像系统。
We all know that visual inspection system has a prominent contribution in defect detection, but does it have the same effect for those parts with high light and reflection? According to Hongbao's vision inspection experience for many years, it is found that the automatic imaging system with novel lighting and imaging technology is needed in the inspection of bright and reflective parts.
生活中常见的汽车配件、浴室设备等部件通常都镀铬,表面光滑且反光强,像水龙头之类的物品必须在出厂时没有任何可见的缺陷。一种新型的自动化视觉检测系统,该系统使用结构化的照明技术、工业机器人和高分辨率摄像机。在系统的设计中,所有的主要组件包括:机器人、照明、摄像机、软件和人机界面。
Common auto parts, bathroom equipment and other parts are usually chrome plated, smooth surface and strong reflection, such as faucets and other items must be delivered without any visible defects. A new automated visual inspection system using structured lighting technology, industrial robots and high-resolution cameras. In the design of the system, all the main components include: robot, lighting, camera, software and human-computer interface.
无论生产过程是怎么样的,最小的缺陷尺寸可能需要检测到0.03 mm2。对于人类操作员来说,识别一个缺陷是0.03 mm2还是0.04mm2,以及每平方厘米是否会出现一个或两个以上的缺陷是困难的。
Regardless of the production process, the minimum defect size may need to be detected to 0.03 mm2. It is difficult for human operators to identify whether a defect is 0.03 mm2 or 0.04 mm2, and whether one or more defects per square centimeter will occur.
为了解决这个问题,我们会用遮盖式立体偏转仪(CSD)系统来自动检测这些零件的高光泽表面的视觉质量。
In order to solve this problem, we will use the covered stereo deflectometer (CSD) system to automatically detect the visual quality of the high gloss surface of these parts.

零件被从托盘和放置在一个180°反射穹顶投光器内,投影仪给机器投上正弦条纹波
为了在某种程度上克服高光物体反光问题,可以使用有源照明系统。在最简单的单结构红光投射中,当物体沿传送带方向向下移动时,光条纹投射在物体上。然后,通过标定的摄像机对物体反射的结构光进行数字化,以确定穿过该物体的线轮廓线,并构建三维表面轮廓线。这样,光可能不会从多个空腔的物体反射到相机,可能需要多个激光器或激光/相机来拍摄,从而增加了系统的成本。
In order to overcome the problem of reflection from high light objects to some extent, active lighting system can be used. In the simplest single structure red light projection, when the object moves downward along the conveyor belt, the light stripe is projected on the object. Then, the structured light reflected by the object is digitized by the calibrated camera to determine the line contour passing through the object, and the three-dimensional surface contour line is constructed. In this way, light may not be reflected from objects in multiple cavities to the camera, and multiple lasers or lasers / cameras may be required to shoot, thus increasing the cost of the system.

计算出被测镜面的斜率后,可以重建镜面的三维形状,并对可能存在的缺陷进行测量分析
相位测量
phase measurement
尽管被动成像系统和主动成像系统都可以确定三维物体的表面特征,但高镜面成像问题仍然是一个难题,特别是对于具有角表面的零件。为了克服这个问题,可以使用一种称为相位测量偏转仪(PMD)的技术。
Although both passive and active imaging systems can determine the surface features of 3D objects, high specular imaging is still a difficult problem, especially for parts with angular surfaces. To overcome this problem, a technique called phase measuring deflectometer (PMD) can be used.
在这种方法中,一些条纹图案被投影到被检测物体的表面。
In this method, some fringe patterns are projected onto the surface of the detected object.

CSD检测镀铬和油漆中的所有表面缺陷类型
使用三台投影仪投影条纹到光滑的物体表面,从而在物体的光滑表面创建纹理。
Three projectors are used to project fringes onto smooth surfaces to create textures on smooth surfaces.
所生成的条纹图案会被反射到零件的镜面上,并随着被测表面和调制后的条纹图案的斜率变化而出现变形。在CSD系统中,这些图像由来自的3台4096 x3000、摄像机拍摄的。这样,零件表面的完整轮廓被照亮,相机可以拍摄到反射的条纹图案。
The generated fringe pattern will be reflected on the mirror surface of the part, and will deform with the slope of the measured surface and the modulated fringe pattern. In the CSD system, these images are taken by three 4096x3000, cameras from. In this way, the complete contour of the part surface is illuminated, and the camera can capture the reflected stripe pattern.

缺陷显示在软件中,它不需要任何编程技能进行设置
这些软件用于可视化任何缺陷,如可能出现的小丘疹、划痕和凹痕(上图,右)。在这个例子中,显示了一个大约0.03 mm2的小“丘疹”缺陷。对于每个部分,任何这样的缺陷都按照缺陷的类型和数量进行分类。系统可以在检查部件、部件类型和缺陷时存储它们的完整记录。通过使用图形用户界面设置特定的参数,操作者可以确定每个部分可接受的级别。通过这种方式对零件进行检验后,机器人将好的零件放入生产线上的垃圾箱中,检验不合格的零件放入废品箱中。
The software is used to visualize any defects such as possible small papules, scratches and dents (above, right). In this example, a small "papule" defect of about 0.03 mm2 is shown. For each part, any such defects are classified according to the type and number of defects. The system can store a complete record of components, component types, and defects as they are inspected. By using the GUI to set specific parameters, the operator can determine the acceptable level for each part. After the parts are inspected in this way, the robot puts the good parts into the dustbin of the production line, and the unqualified parts are put into the waste bin.
视觉检测系统