Flow Rack material management using vision

Hello guys!

Nowadays I am trying to develop a Flow Rack manegament using vision.

Firstly using “Jigs” to identify the boxes and know what material pieces each box has, I want to control how many pieces I have in stock.
Secondly using the change detector I want to know if the boxes are in the right position and if they are there at all.
Does anyone knows if that is possible to achieve?


Fernando Oshiro

1 Like

Hi Fernando
Tracking boxes should be easily implemented with the Jigs feature.
Just stick a marker on each box and the camera could track them. You will need to “Train” the jig detector to give each Jig a name that’s meaningful to your setup.

You won’t need to use change detector to tell if a jig (box) is in its place - the region detection works right away for Jigs!
Just set up the regions of interest in Camera Configuration and enable your trained jig detector (toggle button) in them.
In your application you can now write triggers that respond to a box entering or exiting a region.

The camera configuraion screen will be very helpful in dragging the regions into place on top of the view from the camera.

Note that you can use a plain color camera for this use case, and don’t have to purchase a range (depth) camera.

Let us know how it goes!

Hello Royshilkrot,

Is there any limitation related on how far the camera can recognise and distinguish the JIGS, and if increasing the JIGS sizes, or change the actual camera for a better one would solve it?
One more thing, is it possible to use the notebook’s ou cellphone’s camera to use VISION?
I am trying to configure those, but sadly my instance is not recognising them.


There is not a limitation per se, except the jig markers need to be big enough from the point of view of the camera.
Either the camera needs to be close enough, or if the camera is farther away - the printed markers should be printed in a large format.
The best way to know if simply to print a marker and place it on an object, and then moving the camera.
In the Camera Configuration screen, after you’ve enabled the Jig Detector, you will see a blue marking over the marker if it’s successfully picked up. If it’s too far away - the blue marking will not appear.

In regards to the built-in camera - for most PCs the built-in webcam should work.
However, we do not support Vision on mobile devices yet (unless they are Windows 10 64bit devices), such as Android or iOS. Mind, non-Vision Tulip is available on Android, and can use the phone’s camera for barcode scanning.


Dear @Oshiro and @royshilkrot . I work with Oshiro in Konitech. I am not too familiar with JIGS as him, but I have a question. Is it possible to translate the JIGs to a specific SKU. Putting more clearly: can I have a tag that can be uniquely correlated to a specific SKU (material code) ? Thanks

@flmaeda hi!
The jig markers can give you up to 1000 unique IDs. If that matches the amount of SKUs you have in your system then - yes it can support a 1-to-1 mapping.

To do the actual mapping I suggest using a Tulip Table which has on one column the “Jig ID” or “Jig Name”, and in another column the SKU.
When a jig trigger event comes in you can correlate the Jig ID/Name to the SKU and act accordingly, e.g. show some semantic information.

Does that help clear things up?

@royshilkrot . thank for the explanation. Is the 1000 IDs a temporary or permanente limitation for the Jig markers?

This is our current limitation for Jig markers.
If you have a need for a very large set - I suggest using barcodes or QR codes. Those can be read by another camera or a dedicated barcode reader (handheld or stationary). Tulip Player supports reading barcodes with regular cameras, by using the camera widget in the App builder.

@royshilkrot , do you know if there is a limit of how many jigs a camera can track simmultaniously?

@Oshiro Any limitation on number of concurrent jigs will be imposed by the hardware running the Vision software. With a more powerful PC you could track more jigs.

What hardware are you using to run the Tulip Vision workstation? (in terms of CPU type, e.g. i5 or i7, amount of RAM in Gb, and type of storage, e.g. SSD or HDD)

In case anybody’s wondering about Jig size and how far away they can be read without background noise.


this is great @DewyWCI, thanks for sharing.

I’d be really interested to learn more about what you’ve been building with Vision!! is this something you could share on Show and tell - Tulip Community??

thanks again for sharing this!!


We’ve been doing a lot of testing with Tulip Vision lately. We will be using jig detection on assembly fixtures to know when a build is being performed, and then track runtime, downtime, efficiency, etc alongside operator inputs.
We’ll also be implementing an OCR cell to read serial numbers off of labels and retain the data as well as read results off of a tester in the cell.

Ill add this table to show an tell as well!