Machine vision systems present a cost-effective solution for orientation of complex parts which do not have enough mechanical properties to be used for conventional orientation methods(e.g vibratory bowl feeder with orienting devices for removal of unnecessary part positions).
Main applications of machine vision systems:
- Orienting of parts by a certain feature – slot, label, color, logo, etc.
- Sorting of parts
- Parts quality control – dimension checking, detection of missing features, etc..
- Navigation of industrial robots
Machine vision systems offer 100% control. Only the parts recognized by the vision system are let through. All the rest are removed, even if they tend to be oriented properly.
There are different devices to solve a specific task. Machine vision systems differ in capabilities, image processing speed, color support, communication protocols with third party devices, enclosure type, protection class and of course – they differ in price. The proper functioning of the system depends on the proper choice for machine vision system.
The devices for machine vision inspection are mainly divided into three groups:
- Vision sensors – they are the most basic kind. Usually, the software has simple tools – pattern recognition, brightness, contrast. Their application is in project where proximity switches are not capable to solve the task. In most cases, lighting and the electronic control are integrated inside the device. When a trigger signal has been received, the sensor compares the current image to the stored one during the teaching process. The result from the inspection is “yes” or “no”. The software application provides settings for the I/O signals. This way, the signal from the camera could be brought to a programmable logic controller for further control. Communication capabilities are manufacturer-specific. The price of vision sensors is relatively low compared to the other machine vision devices.
- Vision cameras – a higher class devices, dedicated to more sophisticated inspection projects with multiple parameters. Lighting is usually ordered separately, based on the inspected parts. Similarly to vision sensors, image processing is handled inside the device and the inspection results are provided via the I/O signals. The communication capabilities are extended. Machine vision cameras are used to navigate industrial robots. The price is much higher compared to vision sensors.
- PC-based machine vision systems – the most advanced vision systems. In many cases the hardware(video camera) and the software are independent. This means cameras and software for vision inspection from different manufacturers could be used together in a single project thanks to several integration standards.
Examples for machine vision systems:
Machine vision system for inspection of dice:
Machine vision system for inspection of stick-on balance weights (according to the one-side Fe5 emboss) :