Increasing competition and rising quality standards are prompting many companies to continuously develop their production processes. The worsening shortage of skilled workers in the ageing societies of the West is creating additional pressure to adapt. More and more companies are responding to this by automating parts of their production lines. Modern technologies now make it possible to delegate even more complex tasks to robots, where humans were previously superior to any machine thanks to their five senses. Vision systems combine hardware and software and support devices and systems by capturing and processing images. The strong arm of the robot is supplemented by an artificial eye that is able to recognize objects and structures and handle parts precisely and appropriately depending on their properties. Due to their cost intensity and complexity, such systems were virtually impossible for SMEs to implement in the past. Thanks to the further development of the technology, they now also offer SMEs an affordable and easy-to-implement alternative to manual work.

The basis: a robot arm with peripheral device

The basis of every automated application with a vision system is a robot arm with a suitable peripheral device. The associated image processing system, which is usually software-based, captures images using one or more cameras and digitizes the data. It searches for previously defined features, such as the surface structure of an object, to which the robot responds with a pre-programmed action - for example, by transferring a part to the next work step or rejecting faulty workpieces. An important prerequisite for the functionality of such a system is the right lighting. This ensures that the camera system can correctly detect the key features.

An integral part of automated applications with a vision system is a robot equipped with a suitable end effector. © Universal Robots

Depending on the application scenario, the image material can be recorded and processed in both 2D and 3D. Vision systems with 2D image processing are able to calculate a two-dimensional data set. Objects are captured on the basis of their contours. In this way, the seeing cobot can recognize the position, size or orientation of a part. This is ideal for applications in which objects, geometries or patterns need to be analyzed. However, more demanding tasks such as assembly work or feeding machines with unsorted parts require more complex 3D images. Appropriate sensors enable robots to see in three dimensions. This capability comes into play, for example, when inspecting surfaces, measuring volumes or in the field of quality assurance.

The Austrian medium-sized company Jenny Waltle, a specialist in aluminum, metal and plastic parts based in Vorarlberg, also relies on collaborative robots that are connected to a 3D image processing system. The 50-man company was particularly struggling with the lack of qualified personnel. "In order to continue growing despite this, we had to automate," explains Daniel Waltle, one of two managing directors. For over a year now, two UR5 cobots from Universal Robots have been loading a CNC milling machine in close proximity to their human colleagues.