3D Vision Applications

Automate more robotic guidance applications with less effort

Are you looking to integrate 3D vision into your standard robotic solution? Learn about four types of applications where Pickit 3D vision systems are utilized.

1. Semi-structured bin picking.

Use your robot to pick parts from somewhat structured bins or pallets to load your production line with a 3D vision system. A robot needs eyes to find the positions and orientation of parts on a tray, bin, or in a box. Parts could be separated by intermediate layers, stacked on top of each other, or singulated (on a conveyor, for example).

With layers.

A robot can unload semi-structured parts from a bin, as well as removing in-between layers. Typical application example:

Driveshafts loading.

At GKN, specific parts of a car gear-box need to pass through up to fifty machines to complete the manufacturing process. This means the parts have to be transported internally from one machine to the other. The driveshafts are placed into a bin by a robot in a structured way with intermediate cardboard layers. Placing the parts in the bins can be done blindly. However, for taking them out, GKN needed vision. Why? During the internal transportation, about 10% of the parts would fall over or change position, making it impossible for the robot to pick them from their initial position. Pickit 3D vision systems enabled the robot to detect, take out, and place parts with orientation; not only the driveshafts but the layers in between as well.

Driveshafts semi-structured robotic picking with Pickit and Fanuc at GKN.

No layers.

Parts are already in some sort of order, like stacked or singulated on a conveyor or a table. Typical application example:

Crankshaft, camshaft, head bolt loading.

A robot picks crankshafts from the bin using the Pickit L-HD vision system.

Sheet metal picking.

A robot picks parts from a stack on the table using the Pickit M-HD vision system.

Read our Guide to bin picking in automotive to see more examples from the automotive and railway industries.

2. Depalletizing.

Adding vision to depalletization robot cells increases flexibility while maintaining performance.

Typical applications examples:

Bags.

Animal feed production is the core business of the Dutch company Van Tuijl. It’s not surprising that thousands of heavy bags need to be handled every day. Before automating with Pickit, the unloading of pallets with bags varying in weight and size was done manually, in a non-ergonomic way.

Pickit has a vision engine that finds bags in stacked layers on pallets. The solution has been proven on 97% of all pallets we ever tested. The Pickit Bags engine was designed for depalletizing bags, arranged in a pattern. This engine supports various bag patterns and allows automating the challenging and non-ergonomic task of picking heavy bags from a pallet.

Boxes.

Another example is depalletizing boxes with the Pickit L camera.

3. Random bin picking.

Have your robot pick parts from bulk and load your production line using a 3D vision system. A robot needs eyes to find the positions and orientation of parts on a tray, bin, or in a box. Typical application examples:

Axles / shafts / billets bin picking.

Jinmyung Powertech produces power trains, the core component of heavy equipment, agricultural and industrial machinery. The company improved the machine loading of the CNC machine for shaft production using Pickit 3D.

This application required placing the parts with high accuracy and was solved in two steps. First, the Pickit 3D vision system detects the part, and next to that determines the best pick point of the part without collision to other objects next to the gripper. All thanks to our multiple pick points and tool modelling features. In the second step, the gripper places the shaft on a fixture, where Pickit checks the part orientation. Pickit also finds the empty slot in a tray to place the manufactured part after machining.

Have a billet handling application? Check out our Billet picker solution.

Bin picking of other cylinders: crews, pipes, rods, bottles, sticks, drills, caps.

All other axisymmetrical parts that don’t fit into the “axles/shafts/billets” category are also typical for bin picking.

CV joints bin picking.
Stretch film rolls bin picking.

Sheet metal bin picking

Sheet metal parts would arrive in bins for the next manufacturing step, and the operator at Stantraek had to stage them manually in front of the robot. Pickit allowed them to eliminate the staging so that the cobot could take the parts directly from the bin and feed it into the CNC punching machine.

4. Part localizing.

Guide a robot to a 3D point on a part.

Assembling or installing parts becomes more reliable with the 3D vision for your robot. The robot needs to know your product's position and orientation before starting the automated assembly tasks. Typical applications examples:

Vehicle frame assembly.

Guide a robot to an offset position. with respect to a 3D part

Oftentimes the robot is unable to reliably perform tasks on objects that have no fixed location in space. You need a vision system to locate the object and find the correct location to allow the robot to perform its tasks.

Leak detection, gas sensing.

In this application you will see condensers filled with gas, so making sure there is no gas leaking from the pipes is crucial. Pickit locates the pipes. The robot then knows where to sniff using the gas detector to check for any leaks.

Guide a robot with a motion over a part.

Pickit 3D helps your robot to grind, glue, and deburr parts in your production line. Even when the parts have no fixed position when presented to the robot. This application is the most challenging among other part localizing tasks. Typical applications examples:

Polishing, deburring, sanding, gluing, grinding.

Doosan Industrial Vehicle utilizes Pickit as eyes for a polishing robot (Yaskawa) to detect a vehicle counterweight on a conveyor. The system is used to produce 60 parts per day, out of which 41 are unique in shape. Parts are produced in random order. Previously the operator had to visually identify each part on a pallet, and press the correct button out of the 41 products which corresponds to the product file. This work was a burden for the operator because they had to remember the shapes of all 41 different parts.

Now the robot knows which part is arriving for a polishing step and is guided to perform the correct polishing moves.

Get a quote and info

Looking for a price, production references, or have questions? Please contact us for more information on our vision systems.

Contact us