Non-barcode automated checkout that identifies fresh fruit
Is it possible to identify fresh fruit within plastic bags? This question was investigated by experts from Wageningen Food & Biobased Research. The outcomes provide leads for automated, barcode-less checkout at supermarkets and opportunities for the development of new sustainable packaging.
Self-scan checkouts are increasingly popular among consumers and retailers. They are often much quicker for people buying only a few products, and retailers need fewer cashiers. However, with large amounts of groceries, self-scanning can be laborious. Consumers with full trolleys therefore still prefer a traditional checkout with a cashier.
Fully automated barcode-less checkout, based on advanced machine image-recognition technology, could be the answer. For pre-packed products the task of image recognition is relatively well addressed. However, detection of fresh produce inside semi-transparent plastic packaging is more challenging, as the plastic makes accurate imaging difficult.
Identifying and counting fruit
In the project Optimization of existing MatchX Technology (2019-2020), carried out for self-check-out technology manufacturer ITAB, Wageningen experts met the challenge. They combined near-infrared hyperspectral imaging with spectral orthogonalization; a unique data-processing method that virtually removes the plastic bag from the products.
They successfully validated this new approach for three popular types of fruit: bananas, apples and pears. The equipment was able to identify the fruits and count the number of pieces in three types of semi-transparent packaging that are commonly used in Dutch supermarkets: low, intermediate and high density poly ethylene.
How do new packaging materials impact product quality? Non-invasive measurement of the quality of a product inside a package is very challenging. Methods to achieve this can, for example, help understand the efficacy of several types of new sustainable packaging in maintaining product quality and extending the shelf-life. Near-infrared hyperspectral imaging and spectral orthogonalization could have potential in the development of sustainable packaging solutions. Imagine materials that provide consumers information on the quality and shelf-life of the fresh produce inside, thereby preventing food waste.
As far as conventional plastic packaging goes, the insights from the research allow ITAB to take well-considered decisions about which technologies to invest in for the development of next-generation check-out systems.