Hacker News new | ask | show | jobs
PiScan: An Open-Source Version of the Amazon Dash Button Using a Raspberry Pi (denis.papathanasiou.org)
37 points by dpapathanasiou 4033 days ago
9 comments

A Pi is ridiculously overkill for something like this. There are much more suitable modules. If you're willing to have a bluetooth gateway device you could do it pretty damn easily with an nRF51822 module.
This is great. It would be better if it were battery powered and smaller, like the dash.

This option, using the ESP8266 instead of the Raspberry Pi, looks promising: http://hackaday.com/2015/05/13/an-amazon-dash-like-button-fo...

Thanks!

My initial vision was a single enclosure to house everything, in a small handheld package.

But I just wanted to get all the parts working first.

I'll take a look at the ESP8266; thanks for mentioning it!

A version with a camera would be nice, if it could recognize the packaging and you didn't have to aim at the barcode. A cheap android phone might be a good platform for this (camera, CPU, WLAN, battery all already integrated)
I did think about doing this as a phone app, since there are libraries for converting pictures of barcodes into the corresponding numbers.

The rub is you need a fairly accurate camera, and while it would probably be ok in iOS, the experience in android would vary depending on your hardware.

The dedicated laser scanner is cheap, accurate and reasonably compact (only the head is important, the rest of the handle is cabling and empty space), which is why I'd like to try combining it into a small handheld device, ultimately.

I agree this would better packaged as a phone app. There are several apps out there that scan barcodes to detect a product. Amazon's app is an example. There is a price comparison website in Brazil (buscapé) that also has an app which does that. I don't believe the camera needs to be that accurate, as I have used them on several Android devices, from lower- to higher-end phones.
One such library is ZXing: https://github.com/zxing/zxing
For using the barcode, a dedicated scanner is clearly better (especially faster). Reading the barcode with the camera is reliable as a fallback, but many products have packaging that should be possible to recognize directly with a camera, especially for items you commonly buy.

I don't know, maybe that's overkill and just pointing the barcode at a dedicated reader is actually the better experience.

I thought about doing something like this myself a while back. I use Ocado[1], who have their own scan and shop app. I could possibly put this onto an older iPhone with a battery pack and some sticky velcro and run it from that.

I like what the author has done though, and it's good that the server was open sourced.

[1] - http://www.ocadogroup.com/news-and-media/news-centre/2014/20...

This is an excellent project, nothing complex with the exception of the everyday "USB Laser Scan Barcode Scanner". As for using the Raspberrypi, @IshKebab sure it's overkill. The Pi is a great general way to quickly prototype, lots of different things quickly.

"The Open Product Data (POD) project is a promising start, but its catalog is limited, and they haven’t published an update since early 2014."

Following through the POD project, how many other types of data that need to be opened-up?

Unfortunately the Pi, will draw a hell of a lot of current which would make running on batteries difficult, the esp8266 (an MCU with Wifi support) apparently draws at peak 300mA @ 3.3V.

I've just been playing about with nodemcu (which is eLua ported to the esp8266) on them, whilst also seeing how long it will run for on 2x 18650 3400mAh batteries.

There is another group looking to open up food data here: http://openfoodfacts.org/
It's a neat and useful product. Yes , it would be great if it was battery powered small handheld device or if it could connect to a smart phone.