Most vision-guided robots (VGR) today leverage 2D vision systems. These vision systems allow the robot to maintain a greater degree of accuracy by locating a part in space and ensuring it’s in the correct position. While this is a significant improvement over robotic systems with no vision capabilities, recent developments in 3D vision for robotics are taking identification to an entirely new level.
3D vision offers robotic systems far greater flexibility and reliability in locating a part’s location and orientation in space, even in a random pile of mixed parts.
How Robotic 3D Vision for Identification Works
3D vision systems for robots are highly complex but can follow the same general workflow. First, a variety of patterns are projected onto randomly piled parts, allowing the vision system to measure the distance between the parts and the sensor. Then, part positioning and orientation are recognized by referencing a pattern dictionary and 3D CAD models already registered in the system. Finally, the system determines if the robot is able to grasp the part and sends this data to the robot controller, and ultimately the robot arm.
While not all 3D vision systems for robotic applications work the same, the system workflow described above represents a common method for recognizing parts and transferring this information to a robot.
The Benefits Robotic 3D Vision for Identification
The most immediate benefit of 3D vision for robotics is that it can eliminate any bottlenecks created by the process of manually loading and unloading parts in a robotic workcell. Without the ability to perform random part picking, this manual task is slow and limits the productivity of other robotic systems. Pick and place robots with 3D vision systems unlock the full productivity potential of automation.
In addition to this, robots with advanced 3D vision systems are far more flexible. They can handle a wider variety of parts and account for more variables in their environment, helping maintain high levels of uptime and accelerating ROI. With extensive pre-programming, they also reduce integration requirements, further improving ROI.
Robotic 3D vision systems are only just beginning to be adopted by manufacturers, but the difference this technology makes is clear. Robots with 3D vision system are more flexible, more capable, and more productive.
As 3D vision experiences wider adoption in the coming years, other technologies such as deep learning will see a simultaneous increase in adoption, resulting in drastically more productive and capable industrial robots.
To learn more about the future of identification applications, register for our free webinar, “A New Age of Identification: Deep Learning, 3D Vision, and More.”