Engineer Paco Garcia came up with the idea of building his own pick and place machine after he was inspired during a visit to a plastic injection molding plant and observed one of their automated picking robots in action. His system for toys is a great way to get kids interested in electronics projects and helps instill the importance of putting away their toys, and what better way than using a robotic arm.
Paco’s machine is an open source project, and involves the combination of a robotic arm and computer vision, allowing it to identify and pick up various small, lightweight objects- in this case, tiny toys. The platform itself is a gantry-based system built on an aluminum frame made of 3D-printed parts and uses servos to drive the carriage and arm, as well as operate the gripper.
On the computer vision end, Paco’s pick and place machine is outfitted with a standard camera and a Raspberry Pi 3 B+, which uses OpenCV for object recognition. The Raspberry Pi then hands off commands to an Arduino that controls the servos to actuate the robotic arm. Paco found his build challenging, with the hardest part learning how to generate real-world coordinates using a single camera to grab data in OpenCV, after overcoming that hurdle, the rest of the project came together with excellent results.
As with any great engineer, Paco documented his build process for anyone looking to recreate his pick and place machine and has uploaded all the necessary documentation and software links to his Medium page.