A machine vision system is based on a technology that uses imaging-based inspection for a variety of applications like robot guidance, process control, and automatic inspection, just to name a few. Actually, machine vision involves a lot of technologies, expertise, methods, actions, integrated systems, hardware products, and software to name a few. Let’s find out more about it.
There is a difference between computer vision and machine vision. Basically, this system tries to create an integration between current technologies in fresh ways and implement them in order to resolve real world problems.
The machine vision process involves planning the project details and then finding solutions. And the process begins with imaging and moves towards the automated analysis of the required information like images.
Imaging based Inspection
Primarily, the visual inspection machines are used for image-based inspection and robot guidance. The first step involves the image acquisition, which is taken using lighting, lenses and cameras.
MV software programs are used in combination with digital image processing for the extraction of required information.
The typical parts of an automated inspection system include output devices, software, image processors, camera and lighting.
The imaging device can be part of the unit or may work independently. If it is part of the machine, it is known as the smart sensor or smart camera. When it’s used as a separate unit, the connection is made to the intermediate hardware, a frame grabber or a processing appliance.
Although traditional imaging is also used in machine vision systems, other alternatives are also quite common, such as x-ray imaging, 3D imaging, line scan imaging, hyper spectral imaging, and multispectral imaging.
While most of the visual inspection systems use 2D imaging, many of them also use 3D imaging. In fact, the 3D imaging-based systems are getting more popular with time. The most common 3D imaging method involves triangulation, which is based on scanning.
In this scanning, a laser is shot onto the object surface and seen from a quite different angle. Moreover, in the case of a visual inspection machine, you can achieve this through a scanning motion of the camera.
Lines from more than one scan are combined into a point cloud or depth map. In special cases, stereoscopic vision is implemented, especially when unique features are found in both views of cameras.
Other visual inspection systems are grid or time of light based. For instance, the grid array oriented system makes use of pseudorandom structured system of light.
Once an image is captured, it goes through processing. Generally, the processing is done in more than one stages. Typically, the sequence starts with many tools like filters that change the image. Afterwards, the extraction of data and objects is done. Next, the data is communicated and compared against different target values for creating the desired results.