Better Performance in Machine Construction with Vision Sensors

Pepperl+Fuchs: An increase in performance in machine construction has always been closely connected with the development of sensor technology. Control and regulation technology can only be optimized with ever more precise and faster monitoring via suitable sensors. For a few years now, vision technology has been gaining ground. If at the start, industrial vision was primarily used for quality control, there are now more and more applications where vision sensors are also used directly in machine control. We would like to illustrate this trend with three examples.

PCV Data Matrix Positioning System

In Data Matrix code (2D code), information is encoded on a square area as a pattern of dots. Compared to one-dimensional barcodes (1D code), the information density per unit area is clearly increased and there is therefore a significantly increased error redundancy.

Individual Data Matrix codes can be positioned horizontally or vertically along a path which can be read and decoded with a camera and the exact position output in x and y directions. The process is characterized by the highest data integrity due to intelligent error correction. The code strip is very narrow and is therefore suitable for the narrowest mounting points, smallest curve radii and inclines and declines. Because a minimum of 6 codes are always read, there is a high code redundancy. Even in cases of severe soiling or damage, the exact position is reliably determined, meeting the strictest of requirements.

The PCV reading head uses modern camera technology to detect information (Data Matrix code) two-dimensionally and software evaluates the image, i.e. report the position. The elimination of laser diodes and mechanical parts guarantees extreme robustness and a long service life. Distances up to 10 km including height measurement, extensive diagnostic information and DC switch outputs are just a few features of the unique Data Matrix positioning system.

Laser light sensors

A very different application of vision technology is the laser light principle in the so-called LineRunner. With integrated line-type laser illumination in the sensor, height contours become optically visible, and then 3D information with the same sensor hardware platform that is used in the other examples of this article.

An example of application for the LineRunner in automotive production is the precise installation of doors and attachments. If previously gap and clearance dimensions were checked manually with plastic gauges in production, such measurements can now be detected automatically and without contact using the LineRunner and applied as control variables for robot-assisted mounting.

For gap calculation it is necessary to consolidate the individual distance values detected by the LineRunner with mathematical methods of compensation so that they result in geometrically interpretable values. This is achieved, for example, by adapting the measurement points to known model contours (e.g. circle, radius, tangent). With component edges in an indentation, two sensors are interconnected as a so-called “double head” and the corresponding gap dimensions are detected by both sensors together. Even if a majority of the necessary signal evaluations take place in the sensor, it is clear that significant system integration capabilities are necessary for the task described. This is precisely what VMT Bildverarbeitungssysteme GmbH specializes in.

These mounting strategies, which among experts are referred to as Best-Fit, are already used in modern plants of the Daimler corporation. This technology is characterized by the fact that the mounting robots are directed by sensors to an end position via continual gap and flush measurement which takes place at several locations on the mounting parts simultaneously. The sensors are mounted directly on the gripping tool. Depending on the application, between 6 and 18 LineRunners are used per robot. Using the Best-Fit procedure described, tolerances of the components are balanced as best as possible, and the attachments mounted in optimum position. Overall, the error or tolerance chain between the attachment and body is reduced to a minimum. The example illustrates that intelligent vision sensors such as the LineRunner are becoming more important as components for complex mounting systems and there is no substitute in the system business for professional system competence, which begins with system planning and does not end with 24/7 service.

Vision sensors in the graphics industry

An interesting area of use for vision sensors is misfed sheet control in the post-press area on binders, gather-stitchers or folding machines. Readers are irritated when they find a page missing from the novel they are reading. However, an incorrect instruction leaflet for medication can be a much more serious problem. To prevent this, it is essential that the correct sheet sequence should be monitored on collating and folding machines. Systems previously established on the market work with barcodes on the printed sheets and therefore require additional outlay. Furthermore, the misfed sheet detection option was associated with high costs in the past.

Pepperl+Fuchs has developed specialized vision sensors that, from the point of view of cost, permit misfed sheet control to be integrated in original equipment of machines as standard. The VOS410-BIS and BIS510 vision sensors are compact sensors, that combine camera and evaluation and therefore permit simple mechanical integration.

The vision sensors monitor the sheet sequence on the basis of the print image. The first sheet is automatically “taught-in” and then compared with the subsequent sheets. The sensors can implement this control at maximum speeds of up to 6 m/s and 10 sheets/s. Because of the high resolution, the vision sensors can detect even the smallest differences in the print image.

The vision sensor is simply integrated in the machine’s control unit with an Ethernet interface. Because of its high band-width, this high-performance interface enables fast transfer of images, e.g. of incorrect sheets, for systematic error analysis.

Vision sensors gaining ground

These three examples show how vision sensors increase the performance of machines and plants. It is also increasingly shown that this functions best with sensors optimized for the relevant application. The technical basis of all three sensors is the same, but only with application-specific firmware is the vision sensor a problem solver in machine construction.