New applications for vision systems in robot guidance and quality assurance

Citation

(2006), "New applications for vision systems in robot guidance and quality assurance", Assembly Automation, Vol. 26 No. 2. https://doi.org/10.1108/aa.2006.03326baf.001

New applications for vision systems in robot guidance and quality assurance

New applications for vision systems in robot guidance and quality assurance

Keywords: Machine vision, Automotive industry

In one year, over 59 million automobiles are built worldwide. Automobile manufacturers are under growing pressure from global competition. Accordingly, a high level of flexibility in production is becoming an ever more important criterion for sustainable market success. This high level of flexibility can only be accomplished at large lot sizes through the economical utilization of industrial robots; of which – spread across all production sectors – over 750,000 are already implemented. Image processing systems play a decisive role in robot guidance and quality inspection. With the EASI3D and MONO3D solutions, Darmstadt, Germany’s ISRA VISION is setting new standards for the cost effective implementation of image processing systems (Plates 1-5).

Plate 1 EASI3D – 3D position determination

Plate 2 EASI3D: position and orientation determination, as well as automatic adhesive/ sealant application inspection with one sensor

Plate 3 User guidance and process visualization EASI3D

Plate 4 Crank shaft casing: position and orientation recognition with MONO3D robot vision

Plate 5 3D recognition of sand castings with MONO3D robot vision

To distinguish oneself from competitors, it is necessary to continually offer new, innovative products to automotive manufacturers. Important developments such as just-in-time and “lean” production, which have been continually changing the face of industrial production, are products of the automotive industry. The high flexibility demanded in production in the automotive sector has experts from McKinsey, for example, already talking about a new smart revolution in the automotive industry: In accord with the German Association of the Automotive Industry (VDA), they see the branch in a transition away from a purely functional to a knowledge- and competence-driven division of labor and value-added structure with far reaching consequences for production as well.

Automation technology, in which the automotive industry has been a trendsetter from the beginning, will play a substantial role. Quality assurance and the most precise position assessment for the handling of parts are components of automation technology.

Key to higher productivity

The key to higher productivity, greater efficiency and seamless quality assurance are intelligent image processing systems that ensure the necessary precision and speed in robot assisted assembly. Image processing systems are a fundamental component for more cost effectiveness in automobile production and are heavily implemented, for example, in car body construction – one of the decisive production steps in automobile manufacture.

In the preliminary stages, the various car body components are assembled to form a homogenous basis. Bonding and sealing, a standard process for industrial joining technology, are extensively used. Sealing and support adhesives are now used for the assembly of the roof, side frames, A and B columns, hood, trunk lid, doors and spare tire well. In order to prevent recalls, the process of applying adhesives and sealants must be fully monitored.

One instead of two image processing systems

Until now, it was necessary to use two image processing systems for the various applications. Before applying the adhesive and sealant, one system determined the exact position of the car body component to be joined in three dimensions (3D). The other system inspected the adhesive and sealant bead with regards to its width, position, and the absence of interruptions.

With the EASI3D (Enhanced Adhesive and Sealant Inspection 3D) from ISRA VISION, one image processing system can completely meet the demands of both. The integrated multi-lightband projector makes it possible to not only use holes as features, but also any type of edge, contour or corner on the car body component as a measurement feature.

EASI3D first determines the correct spatial position of the component to be bonded with a precision of up to 0.1 mm. This information is then transmitted to the robot as a corrective vector for the application lane, so that the adhesive or sealant can be applied exactly to the prescribed place. As a further step, EASI3D already monitors the quality of the adhesive or sealant application process during the application, making an inspection afterwards unnecessary. For example, EASI3D can check whether the application is carried out uniformly and without air pockets throughout. This fully automated inspection provides high process reliability.

By combining both 3D measurement technology and in-line quality control in one compact system, not only is flexibility and cost effectiveness improved, but new application areas that result from the high level of cost effectiveness are opened as well.

Higher cost effectiveness with only one camera

MONO3D from ISRA VISION opens the door to new application areas of image processing in robot guidance. This procedure makes it possible to precisely determine all six degrees of freedom (position and orientation) of three dimensional objects from one single captured image with only one camera.

How does this benefit the user? From now on, one camera with the necessary peripherals is enough to supply precise determination of position, where in the past, two or more had to be implemented. This then leads to cost reduction in device configuration, floor space and installation and commissioning. Since there are no constraints on position speed, the implementation of an image processing system has only now come into discussion for many, also simpler production processes such as automated assembly on a moving conveyor, due to economic as well as production and technical considerations. For example, the MONO3D sensor mounted on the robot arm can carry out inspection during the removal of parts from a container or unloading at the highest possible speeds. Stationary implementation of the camera system is naturally also possible.

Both image processing systems – EASI3D und MONO3D – use the latest software developments and provide simple setup and operation within a user-friendly user-interface. The teach-in procedure simplifies programming. Complete documentation makes effective quality management easy. Acquired values are saved in a system database; extensive statistics allow for precise process analyses.

For more details visit the web site: www.isravision.com