Machine vision system is the use of machines to replace the human eye for observation and detection, with automation, high precision, adaptability and other advantages, widely used in industrial manufacturing.Machine vision system includes software system, motion control, optical components and other parts, of which the optical components by the industrial camera, industrial lenses, light source, etc., the three parts are to do the right selection in order to get high-quality imaging, faster completion of the visual inspection.
The main characteristic parameters of the camera are the following:
Resolution: the ability to distinguish light and dark details;
Maximum frame rate: the rate at which images are captured and transmitted;
Image noise: disturbing information that affects image quality;
Target surface: the size of the photoreceptor chip.
The two largest categories of industrial cameras are CCD cameras and CMOS cameras. The main difference is that the sensor transmits data in a different way. ccd cameras have a wider output bandwidth and consume more power, while CMOS industrial cameras are highly integrated and will consume less power.
Industrial lens types are varied, the selection usually need to consider the working distance, phase size, interface, as well as the field of view, resolution and other factors, professional engineers also need to accumulate experience in order to quickly judge the demand, fit the programme content.
The most important task of the light source in a machine vision system is to create a strong enough contrast between the features to be detected and the surrounding surfaces that do not need to be detected, so that the subsequent software algorithms can easily capture the feature values in the image. The selection of the light source usually requires attention to the shape and size of the light source, the colour of the light source, and the angle of the light source. A suitable choice of light source can separate the target features on the object from the other features by high contrast, in order to reduce the difficulty of algorithm design and processing.