Applying Vision with Robots

Aug. 18, 2020
Incorporating vision into existing production operations is cheaper and easier than ever before, but how do you go about selecting the right camera for your robot application?

For many robot applications, vision technology is not required. But there are a number of applications where it can be helpful. For example, picking and placing different objects, handling of multiple object types arriving to the workcell in different orientations, or rapid changeovers when switching between products.

Understanding how to apply vision to robots for these kinds of applications first requires a basic understanding of the different robot camera typesโ€”2D, 3D, and 2.5D.

Kristian Hulgard, general manager of OnRobot Americas, a supplier robot end-of-arm tooling and cameras, explains that 2D camerasย are โ€œthe cheapest of the bunch and though they are reliable, they are the least versatile. Typically, 2D cameras determine length and widthโ€”the X and Y axisโ€”but are unable to determine height, which limits the number of applications they can support.โ€

3D cameras, according to Hulgard, provide all the visual information a robot could possibly need across all three axes and incorporate object rotation. However, this broad functionality comes at a price, says Hulgard, as 3D cameras are โ€œthe most expensive cameras out there and they also have a reputation for being difficult to integrate and operate.โ€

Hulgard contends that 2.5D camerasย are the best option for most robot applications because they occupy a sweet spot between 2D and 3D camerasโ€”both in terms of cost and capabilities. โ€œCapable of determining the height of objects, 2.5D cameras are ideal for scenarios in which objects differ in height and when items need to be stacked,โ€ he says.

In addition, 2.5D cameras are considerably less expensive than 3D cameras and considerably more capable than 2D cameras.

Explaining 2.5D cameras for robot applications from the point of view of OnRobotโ€™s own Eyes vision system, Hulgard says this new 2.5D camera and vision system can be used on โ€œall major collaborative and light industrial robot arms. In particular, Eyes can be mounted externally, which brings increased versatility to your vision setup.โ€

Hulgard says mounting requirements for robot cameras are an important factor to consider as some camera systems are either integrated directly into the robot arm or have to be attached to the robot's wrist to function. He says OnRobotโ€™s Eyes can be mounted anywhere within the production cell.

โ€œExternal mounting eliminates the issue of cables running alongside your robot,โ€ he says. โ€œIt also allows you to optimize cycle time, because Eyes can take care of image capture and processing while your robot is working on another operation. In addition, other camera systems often require time-consuming calibration process; Eyes can be calibrated using just one picture.โ€

About the Author

David Greenfield, editor in chief | Editor in Chief

David Greenfieldย joinedย Automation Worldย in June 2011. Bringing a wealth of industry knowledge and media experience to his position, Davidโ€™s contributions can be found inย AWโ€™sย print and online editions and custom projects. Earlier in his career, David was Editorial Director ofย Design Newsย at UBM Electronics, and prior to joining UBM, he was Editorial Director ofย Control Engineeringย at Reed Business Information, where he also worked onย Manufacturing Business Technologyย as Publisher.ย 

Sponsored Recommendations

Food Production: How SEW-EURODRIVE Drives Excellence

Optimize food production with SEW-EURODRIVEโ€™s hygienic, energy-efficient automation and drive solutions for precision, reliability, and sustainability.

Rock Quarry Implements Ignition to Improve Visibility, Safety & Decision-Making

George Reed, with the help of Factory Technologies, was looking to further automate the processes at its quarries and make Ignition an organization-wide standard.

Water Infrastructure Company Replaces Point-To-Point VPN With MQTT

Goodnight Midstream chose Ignition because it could fulfill several requirements: data mining and business intelligence work on the system backend; powerful Linux-based edge deployments...

The Purdue Model And Ignition

In the automation world, the Purdue Model (also known as the Purdue reference model, Purdue network model, ISA 95, or the Automation Pyramid) is a well-known architectural framework...