Vision-guided robot lends coffee roaster a hand

In manufacturing, a weak link can decrease productivity and increase cost. For a large coffee roaster processing 650,000 bags of beans each year, that weak link was a robot used to unload pallets stacked with bags of raw beans to a conveyor that feeds the roaster. The robot’s gripper arm was tearing the burlap bags and spilling coffee beans on the floor. This caused safety issues, wasted about 100,000 pounds of beans every year and directed expensive labor to clean up duty.

So plant managers turned to an advanced vision-guided robotic in-feed system (Concept Systems) for a helping robotic hand. The new system uses two cameras, two lasers and software to scan the bags, calculate the contours and create a 3D profile of the pallets, which are each loaded with twenty 150-pound bags in four five-bag layers. A computer model is constructed for every tier of bags then run through an algorithm that identifies its unique features and determines the precise position and orientation of each bag.

This information determines the optimal pickup point on the bags and is used to dispatch the robot to each bag for pickup. The robot’s end effector pierces the bags and picks them up using pneumatically operated tines to push the burlap out of the way and penetrate the bag without tearing it.

As each layer is removed, the 3D model is updated for the next layer. To determine which bag to pick up first when presented with a new pallet or a full tier of bags on a partial pallet, the software directs the robot arm to the highest bag.

The vision-guided robotic in-feed system has helped the coffee roasters double its green bean in-feed rate. And, by keeping the coffee beans from spilling all over the floor, the company has cleaned up waste and safety issues and significantly reduced labor costs.