Supermarket shelves can use image recognition cameras to scan the amount of stock on the shelves and alert staff when new products need to be delivered to stores. The technology, which already exists, can help to lead to a more automated shop and processes for those who provide the products.
Image recognition Company Trax, which says it has analysed more than 36 million images in real-time, has created a ‘smart-cooler’ system which can analyse what is on the shelves of a drink cabinet.
“We’ve done a couple of pilots recently where we’ve took the human interaction out of it,” Neil Gowing, Trax’s General Manager, said.
“We’ve put a camera with our image recognition software onto the cooler door, whereby it takes 10 images per hour, for example, so head office is still getting that report back automated and they’re able to look at stock levels, they’re able to look at promotional compliance, and more.
“That’s one thing that we’ve been trialling and we can see a lot of demand for that from the market, particularly across Europe.”
At present the system is planned for manufactures who are providing shops with stock, to allow them to monitor sales, see the items are positioned in the way they planned and to monitor stock levels. However, Gowing said that in the future, the company could work more closely with supermarkets themselves.
In this new setup, which is being tested in cooling cabinets to start with, the system can upload data into the cloud, allowing manufactures away from the store to assess levels of stock not only of their own products but also those of competitors.
As Google proved this week, image recognition tech and the algorithms that power them are not easy to get right.
The company says its set-up can recognise different products based on their size, and when it comes to drinks it can even tell the difference between different products by the colour of the liquid.
As well as testing the cameras in refrigerated cabinets, the company has also been working on capturing what is on the shelves using image recognition by video.
This method allows a member of staff to walk down an aisle and capture all of the products that are on the shelves using a mobile phone. The data is then uploaded to the cloud and can be accessed remotely, as well as via the phone used to take the video.
Gowling said that the approach was an improvement over having to take separate images of different sections of the shelves, as it would allow staff from product manufacturers to get a view of how their products are performing and being displayed.
It said two major manufacturers are testing the video set-up at present and it hopes to have a beta released later this year.