Random 3D bin picking is a developing robotic skill that requires robots to see and act more like humans.
Typically, a robot is instructed to pick up parts at fixed locations, like the escapement of a feeder bowl or the pockets of a thermoformed tray. To save space and cut costs, manufacturers would prefer that robots pick randomly oriented parts out of bins, boxes, and totes.
The 3D Bin Picking Process
Picking from a bin isn’t an easy task for a robot equipped with machine vision. Human dexterity is a skill that is a perfect combination of touch, vision, and hand-eye coordination. People can grab random parts from a bin without much thought, but robots struggle with this task. Give a robot well-arranged items and it can do the job, but toss the items in a bin and the robot struggles to even figure out what it’s seeing.
For a robot to be able to effectively pick random objects from a bin, it requires a point cloud map. A point cloud is a collection of data points that are defined by a given coordinate system. To create a point cloud, a stereo machine vision camera generates a 2D depth map. Each pixel on that map is then re-projected to 3D space. This results in a 3D model that is highly accurate.
Challenges for Random 3D Bin Picking
The challenges are great for random 3D bin picking. Multiple technologies must work together flawlessly so that 3D bin picking is possible. The 3D model of the part, the bin, the robot end effector, the placement target, and environmental obstacles must all be accounted for.
A model of one or more ways to pick up the part and put it in the target location must be created, then image analysis software must be able to locate the part and identify obstacles to removing it from the bin. Finally, path-planning software must find a collision-free route from start to finish. And aside from just picking the object, the robot must be able to navigate to the bin, move its end effector, and properly handle the part.
3D Machine Vision Offers a Solution
Digital imaging, recognition, data processing, and optical technologies serve as the “eyes” and “brain” of industrial robots, allowing them to meet the requirements of high accuracy, speed, and low maintenance.
To learn more about how 3D machine vision is enabling advanced random bin picking, watch our free archived webinar, Realizing the Benefits of 3D Bin Picking, hosted by Canon.