3D Vision-Guided Robotics

Uses a 3D camera to find a workpiece’s position and orientation and guides an industrial robot to pick or process it.

3D vision-guided robotics is the industrial-automation term for this setup: a 3D camera — commonly structured light, stereo, or laser triangulation — is mounted at a robot workstation, captures a point cloud of the workpiece, and recognition plus pose estimation gives each workpiece’s 3D position and orientation; hand-eye calibration then converts this into the robot’s coordinate frame to guide the arm to pick, place, assemble, or polish. Traditional industrial robots rely on teach-and-playback, which requires the workpiece to be in exactly the same spot every time; adding 3D vision lets the robot handle parts that sit in random positions or messy stacks. Typical applications include bin picking (picking loose parts from a bin), mixed-case depalletizing, machine tending, and locating parts for assembly; Chinese vendors in this space include Mech-Mind and Percipio. It can be seen as the mature, factory-grade engineering version of the “perceive, then grasp” pipeline, though the object types and tasks are usually fixed in advance.

ExampleAt a parts factory, a 3D camera photographs metal parts piled randomly in a bin; software matches the point cloud against each part’s CAD model to compute its pose and a graspable point, guiding the robot arm to pick up parts one by one and place them on a machine tool — this is bin picking for machine tending.

Also called

3D Vision-Guided Robot, Vision-Guided Robotics, VGR

Related

Bin Picking · Machine Vision · Structured Light · Hand-Eye Calibration · 6D Object Pose Estimation · Mech-Mind Robotics

Sources

Wikipedia: Machine vision
Wikipedia: Bin picking
Mech-Mind Robotics 官网（3D 相机与视觉引导应用） (Chinese)