Machine vision-based method and system for measuring 3D pose of a part or subassembly of parts

A machine vision-based system for controlling a reprogrammable industrial automation machine to perform a work task on a part or a subassembly of parts which has an unknown pose, the system comprising:

• at least one 3D sensor mounted on an industrial automation machine to move therewith, the machine being programmed to follow a nominal motion path based on a known reference pose of a reference part or subassembly of parts;
• a reference cloud of 3D voxels which represent a reference surface of the reference part or subassembly having the known reference pose;
• a mechanism to move the machine and each mounted 3D sensor together to acquire a sample cloud of 3D voxels which represents a corresponding surface of a sample part or subassembly of the same type as the reference part or subassembly and having an actual pose different from the reference pose;
• at least one processor to process the voxels of the sample and reference clouds utilizing a matching algorithm to determine the pose of the sample part or subassembly;
• and
• control logic to apply a reference-to-sample transform to the nominal motion path to reprogram the machine so that the machine follows a corrected motion path to compensate for the difference between the reference and actual poses.