We are based in NSW and would like to build a DIY Dimension, Weigh & Scan (DWS) station that can capture product dimensions (L×W×H), weight and the barcode.
The requirement is if we place a product in the station, it should be able to capture the dimensions (LxWxH) of the product, the weight of the product and the barcode of the product should be scanned. All this data should be pushed to an API.
We would like some assistance with the products to select. We have come across the below parts but would like your guidance as well.
ToF depth camera - to capture dimensions and possibly the barcode
A load cell - Could you recommend a suitable load cell or weigh scale for our purpose. The product would be around a few hundred grams - 25Kg.
A barcode/qr scanner - if required.
A Raspberry Pi 5 to run everything. We will be building any code thats required plus a UI.
Your guidance on this would be highly apprecieated. Let us know the feasibility of this project.
A very interesting project! We’d love to see where this goes
Before jumping into any recommendations, what kind of products will you have to capture?
Will they all be in boxes? Are you only after the largest dimensions in the 3 axes?
Depending on your approach, these products ought to work, though there may be simpler solutions.
It’s like small to medium items like CDs, vinyl records, t-shirts (rolled up), shoes (in boxes), handbags.
Are you only after the largest dimensions in the 3 axes? - This question I didn’t get. I think the answer is yes because the requirement is to figure out the packaging for the products.
An added bonus would be if we could identify what the product is with some AI magic.
You can proceed this way-the system uses a Raspberry Pi 5 as the controller. A depth camera captures product dimensions. A load cell platform measures product weight. A barcode scanner reads barcodes and QR codes. The Raspberry Pi processes all sensor data locally. OpenCV handles dimension detection and object segmentation. The barcode scanner connects through USB or serial. The load cells connect through an HX711 amplifier.
A depth camera might be okay in that regard, I was originally thinking of a sliding plate that could colide with the product (like the shoe sizing devices - Google is calling it a “Brannock Device”)
To get a barcode you’ll need a standard camera (or scanner), a ToF returns only the distance.
This is great advice. Ya it makes sense to do a POC first.
I guess the main components would be:
A Raspberry PI 5 (How much memory would it require to run a basic image recognition model?)
A ToF camera
A barcode scanner
Makerverse Load Cell Kit
Other components such as cabling, etc.
Please confirm if the above looks right?
I will loop in Devin. Both of us we ll be doing the POC for this project. The next step would be to get a costing for all components. Do you think you would be able to assist with this?
