Our View #5 - 2015 Q2

Vision and robot systems in the laboratory

Written by Jørgen Læssøe

Automatic laboratory systems with integrated vision- and robot technologies are expensive and not very fast. Despite this, there are still great advantages in using a robot and a number of vision systems to systemize measurements and attain reproducible results. In the medical sector there are high demands for quality and many resources are spent on monitoring the production. In this area an automatic measurement system using robots and cameras is a good and reliable solution. Vision systems are usually used in production lines where the quality of every single component must be inspected, but there are also tasks in the laboratory where using vision technology can be an advantage.

Automation of laboratory routines

In the production of plastics components, namely for medical use, the parts are controlled by sampling. To inspect every single part would be a very expensive solution. The mechanics to collect the components and present them in the correct manner to the camera and light would be too costly. This is why statistical methods are used, and samples are taken out at certain intervals. These parts are examined in the laboratory with gauges, profile measuring devices etc. and they are checked for surface quality by simply looking at the surfaces. At the big plastics producers this task is very extensive and demands a large and well educated group of personnel. For a long time there has been a wish to automate, but it has been too difficult to use vision systems. For a vision system to work reliably the components must be fixed and turned with precision to get exact measurements.

The robot helps the vision system

You would normally think that vision systems exist to help robots – but in this case it’s exactly the opposite. In the method JLI vision has developed for a plastics producer, a robot and several vision stations - all optimized for a specific inspection - are used to carry out each control. The robot picks a part from a tray and leads it through the different measuring stations. Every station consists of a light source and a camera connected to the vision system. The robot holds the part in one end and the detail to be examined is held in front of the camera and lights. The part is then turned so the right angle for viewing is achieved. The vision program analyses the picture, makes geometric measurements and examines the surface. By doing this small flash, missing material and dimensional errors can be found. The system is not fast, and it can take up to 20 seconds to collect, position and measure the part in three different inspection stations. It is the robot that is the limiting factor of the process – the vision analysis is usually carried out in fractions of a second. The sequence of analyses is still a lot faster than a manual control would be.

Useful analyses

At the same time as the measurements are carried out, the data is stored. This creates a great potential to work statistically with the production. The analyses can reveal if there is drift in the production and by watching the trend-graphs it is possible to react before the components exceed the tolerances. The statistics can also be used to watch wear on tools and for every batch the quality of the raw materials can be evaluated. Another important point is that the system will give completely uniform measurements, independent of the controller.

A typical measurement sequence

The robot works with such precision that much of the special mechanics that would otherwise be necessary is eliminated. A typical measurement sequence for a plastic cylinder can be as follows:
The robot collects a part from the tray. After this, it is positioned in front of the first camera using line scan technology. The part is rotated while a light source is illuminating the surface. By using a line scan camera, a perfectly unfolded picture is created where the lighting is completely even. Next station is a boroscope with fibre optics. The cylinder is held over the boroscope and turned so the inside of the cylinder can be examined. Several critical details are important for the correct mechanical function and the inspection must be 100% accurate. The last station can be a camera with a dynamic light box where different light and dark fields are used enhance the optical characteristics.

Examining the surface quality

Examination of surfaces for visual defects can be difficult for an operator. Different operators may take different decisions on good and bad parts, whereas a vision system measures details on the surface and does it the same way every time. This is why a vision system will give a uniform judgement of the finish. With these three cameras and programmed light sources you have a general system capable of handling most plastic parts within a maximum physical size. Plastic parts normally carry a form number identifying which cavity the part came from. These numbers are often very difficult to read – also for vision systems, so it is advisable to keep track of the collections of parts, so it is know from which cavity they came from, and where in the magazine they were placed. By doing this the measurements can be related to individual cavities.


1. Unfolded picture of the entire surface including a scratch.

Housing White Mark DIAG 

2. Inside picture with flash.

Cartridge Holder Membrane Entrance Flash DIAG

3. Mark on the end of the case.

Connector pipe 2 Stress mark

4. Boroscope picture of internal details.

Boroscope Image2 

Jørgen Læssøe