Exploring Machine Vision Software in Vision Guided Robot Applications

Vision-guided robotics (VGR) are an exciting synthesis of machine vision and automation technology. A vision-guided robot can respond to its environment using the input from machine vision software. This makes it capable of reacting to changes around it better than a robot that relies on external sensors.

3D Machine Vision Accelerates the Push Toward Industrial Automation

Conventional robotic systems are effective in highly controlled environments. However, not all factory processes can be optimized to meet robots’ needs. Plus, the rise of collaborative robotic applications means systems will increasingly be expected to deal with the most unpredictable element: Humans.

The answer? Sophisticated robotic “sight” using 3D machine vision software.

Today, vision-guided robots are most commonly found picking and placing components in industrial environments. Only a few years ago, such an application was virtually impossible: Robots could only pick and place items virtually identical in size, location, and orientation. Today's software recognizes items in new orientations based on those previously encountered.

However, this is only the beginning. Machine vision experts continue to push forward in the frontier of random bin picking, in which robots are able to manipulate binned items that are in completely natural, unstructured orientations. Recent innovations in machine learning could make it possible to imprint thousands of objects and orientations on a robot's software before deployment.

VGR Can Eliminate Dangerous Tasks Other Robotic Systems Can’t

Picking and placing could save industrial enterprises billions of dollars. However, there are other applications being explored. One example comes from the Odyssey Machine Company of Ohio, utilizing Cognex vision system components.

Odyssey’s VGR system for foundries eliminates hours of sawing and cutting. Instead of multiple professionals working together to cut equipment using dangerous tools, robots handle the most risky tasks.

The system works as follows:

  • The machine vision system identifies the location and type of an infeed part;
  • The robot picks up the part and loads it safely onto a three-stage sprue saw;
  • The robot takes the part from the saw and uses a high-speed cutter to remove flash.

Through implementation of this system, two operators were assigned to more high-value tasks and safety risks were curtailed. The president of Odyssey asserted that the system could pay for itself within six months.

VGR: The Baseline for Robotics in 2017 and Beyond?

For most of robotics' history, only a fraction of robots used vision systems. Thanks to powerful software and inexpensive, compact parts, vision systems may soon become the norm.

Partially due to software, even traditional challenges like calibration are becoming simpler: In the Odyssey Machine Company system, for example, calibration only needs to be performed when the conveyor system itself is moved in relation to the robot, saving countless hours over old-fashioned implementations.