Industry push could see barcodes replaced by QR-style 2D codes within two years

midian182

Posts: 10,634   +141
Staff member
Forward-looking: For almost five decades, the humble barcode has been found on the packaging of almost every consumer item in the world. But there's a push to get rid of them within just two years, replaced by QR-style 2D barcodes.

GS1, the global non-profit organization that develops and maintains standards for supply chain communication, including barcodes, aims for widespread adoption of QR-style codes by 2027 through its Sunrise 2027 initiative.

The 2D versions of barcodes look like squares or rectangles containing many small, individual dots – QR codes are one version – and can hold a lot more information, which is encoded on both the horizontal and vertical axes.

That extra information could be invaluable to stores. Retailers will, for example, be able to immediately respond to product recalls and identify faulty items. In addition to viewing prices, sellers will be able to see when food is nearing its expiry date and offer discounts. The codes can also greatly improve inventory control.

Consumers can benefit from the 2D codes, too. Scanning one could lead to instructions on how to recycle items in compliance with environmental laws. The codes can also give access to loyalty points, games, and coupons.

Something else the QR-style codes can do is offer extra information on ingredients and allergens. That would help those with conditions such as celiac disease – much food packaging fails to disclose the risks of ingredient cross-contamination.

On a practical level, items could have several small 2D codes printed on their packaging, meaning crumples and tears wouldn't prevent products from being scanned, as is often the case with barcodes.

Despite the environmentally friendly recycling information contained in the 2D codes and the fact they have been found to reduce food waste by up to 40%, there are concerns over the greater demands the extra data will place on cloud computing resources. If every product starts coming with a slew of data attached to it, their carbon footprint will increase. There are also the costs of redesigning packaging to consider.

Since they were first used to label car parts in Japanese factories in 1994, the popularity of QR codes has continued to increase, and their use exploded during the pandemic.

Several major US companies are now transitioning or actively supporting the shift from traditional barcodes to QR codes, including Procter & Gamble, The J.M. Smucker Company, and Wegmans Food Markets.

Even if all retailers make the move to 2D codes, barcodes will likely continue to be used in manufacturing and warehousing.

Permalink to story:

 
QR codes stickers need to have double the code images within a single tag as consumers and sellers don't want to jump through hoops figuring out what the product is if a very small part of that code is faded or damaged.
 
I have been using a high density QR codei n a project of mine,2 bits per dot on the grid so it encodes 4 times more information per area. Each dot is in color, pure CMYW, Cyan, Magenta, Yellow and White. I don't use black on the inside to save ink. Black is only used on the outer frame, which also encodes the size. With CRC16 for validation. Currently I use 5x5, encoding 24-bit numbers, which I use as indexes in a DB. The QR codes come in 3 different sizes, 5, 10 and 20 cm on each side. Thus, with the size encoded in the frame, and information from the optical system of the device, the distance to the label can be estimated. With the camera capturing video in 1080p and good lighting (and/or an acceptable sensor) a 20cm tag can be detected and decoded from 10 meters away.
 
Last edited:
QR codes are harder to scan, They're a mess, compared to traditional bar codes. The slightest flex and you gotta try again. This is bad news.
The best QR detection and reading algorithm I've tried is the one from WeChat. OpenCV includes it. Well, it's not included by default, you have to compile the library, adding that module. You can even make it use a small neural network for denoising and super-resolution of the image of the labels, for small QRs, or for conditions that affect the image quality, such as bad lighting or sh1tty sensors. Its results are very good in challenging conditions.
 
QR codes are vastly superior to bar codes, it's not even close. Here's an interesting video from a guy I follow on YT:

 
QR codes are harder to scan, They're a mess, compared to traditional bar codes. The slightest flex and you gotta try again. This is bad news.

It's not hard, they always work for me....guess I'm just lucky or you have a crappy scanner.
 
The best QR detection and reading algorithm I've tried is the one from WeChat. OpenCV includes it. Well, it's not included by default, you have to compile the library, adding that module. You can even make it use a small neural network for denoising and super-resolution of the image of the labels, for small QRs, or for conditions that affect the image quality, such as bad lighting or sh1tty sensors. Its results are very good in challenging conditions.

They also have the best of CCP spying! SMH
 
Back