- By Andrea Pufflerova
- August 10, 2021
- Feature
Summary
Vision-guided robotics deploying 3D machine vision has become the cornerstone of the global race for greater automation. In contrast to 2D vision systems, which can only provide 2-dimensional output images without depth information, 3D vision technologies are able to cope with the intricacies of the 3-dimensional space, delivering a 3D point cloud with X, Y, and Z plane data.
Vision-guided robotics deploying 3D machine vision has become the cornerstone of the global race for greater automation. In contrast to 2D vision systems, which can only provide 2-dimensional output images without depth information, 3D vision technologies are able to cope with the intricacies of the 3-dimensional space, delivering a 3D point cloud with X, Y, and Z plane data.
In practice, this means that while 2D vision systems may be used for simple robotic tasks such as barcode reading, character recognition, or dimension checking, they will not suffice for complex applications that rely on precise depth calculations for accurate, reliable and safe robot navigation. These tasks include robot guidance for object handling, assembly, bin picking, measurement of volume, holes, or other shapes, quality control and other applications.
While there are a number of technologies that enable the capture of 3D scenes, the structured light method is one that can deliver exceptionally high precision and accuracy. Photoneo PhoXi 3D Scanners, for instance, provide a point cloud resolution of up to 3 Million 3D points and an accuracy of 25 - 500 μm across its five models. Each model is suitable for a different scanning range - from 15 centimeters up to 4 meters.
The scanners can handle dark, metal, or shiny materials very well, which are surfaces that generally present a big challenge for 3D vision systems. Dark materials absorb most of the laser light, reflecting only a small proportion of it, while shiny surfaces are likely to reflect the light in a specular way–i.e. in one direction and often outside the camera view, or they reflect the light between individual parts of the scene which causes interreflections. In both cases, the light distortion interferes with a proper data acquisition process.
The scanners can adapt to surface variations, offering a wide spectrum of settings that enable the regulation of exposition or filtering of the acquired data. Thanks to this, they can perform very well with dark and shiny surfaces.
The stable and robust performance of the scanner on these surfaces played a major role in selecting it for a project of robotic picking of heavy metal spheres from a container and inserting them into a measuring machine.
Robotic picking of metal spheres
The system was integrated by Rossum integration, a designer, manufacturer and integrator of robotic workstations and automated lines for manufacturing companies. The customer needed to automate the above described process to make the procedure simpler and more efficient and to reduce physical activity involved in it.
The metal spheres, which are semi-finished products for ball bearings, are randomly placed in a container. The scanner is mounted above the bin and in each cycle makes a scan of the whole container, localizing the sphere which is to be picked. The deployed robotic intelligence system then sends a command to a Fanuc robot equipped with a vacuum gripper to perform the picking action using a suction cup.
Challenges of the project
The project was challenging due to several factors. On the hand was the metal, glossy surface of the spheres. On the other hand, a sphere has an infinite number of rotation invariances so the gripper needs to approach it perpendicular to its surface to avoid the risk of its falling off the gripper. However, by using various types of gripping points, it is possible to set the required number of gripping positions.
Finally, the weight of the spheres also played a role - because they are made of heavy metal material, the task required a vacuum gripper that would be able to reliable handle and transfer the spheres from the bin to a measuring machine.
Benefits of automation
The deployed robotic intelligence system ensures quick localization, fast trajectory planning, and reliable collision avoidance. Combined with the 3D scanner, it provides a fast and robust solution that meets the customer’s requirements.
The final automation solution boosts productivity and efficiency and saves costs. Considering the heavy weight of the metal spheres, it also eliminates the risk of injuries and health detriment.
About The Author
Andrea Pufflerova is a PR specialist at Photoneo, a leading provider of smart automation solutions powered by robotic vision and intelligence. She has a master’s degree from the University of Vienna.
Did you enjoy this great article?
Check out our free e-newsletters to read more great articles..
Subscribe