Bin Picking

Bin picking is the robotic task of locating and picking parts lying randomly in a bin or tote, using machine vision, usually 3D, to find a part and its pose, plan a grasp and plan a collision-free path out of the bin.

A 3D sensor captures the bin contents, and software segments the scene, recognises parts by matching CAD models or using learned detectors, and estimates their position and orientation. A grasp planner selects a reachable grasp, and a path planner checks that the robot and gripper can reach it without hitting the bin walls or other parts. If no grasp is feasible, the system may try another part or disturb the bin to rearrange it.

Bin picking feeds machine tools, presses and assembly lines directly from containers, reducing manual handling or the need for dedicated feeders. In logistics, the related task of piece picking handles mixed items from totes for order fulfilment, where learned grasping methods are widely used.

Difficulties include reflective or dark parts, interlocking shapes such as springs, parts lying against walls and achieving the required cycle time. The last few parts in a bin are often the hardest to pick. Engineers assess pick success rate, cycle time and part damage in trials with representative parts and containers before committing to a design.

Key points

Picks randomly oriented parts from bins using vision
Combines 3D sensing, pose estimation, grasp planning and path planning
Reflective, interlocking and wall-adjacent parts are difficult
Trials with real parts determine success rate and cycle time

Where AiVibe comes in

AiVibe designs and manufactures the AiAmbA AI Factory, whose edge devices and AI agents let people talk to robot controllers in plain language. Robotics perception is an AiAmbA use case, and robot safety functions follow ISO 10218 and never depend on the AI layer.

Explore AiVibe’s work in Robotics & Physical AI →

Related terms

Terms that refer to Bin Picking

Hand-Eye CalibrationRobotics & Physical AI