AI now sorting peaches, melons at supermarkets across Korea

Tech

AI now sorting peaches, melons at supermarkets across Korea

Published: 30 Jul. 2024, 07:00

KIM JU-YEON
kim.juyeon2@joongang.co.kr

Peaches and nectarines sorted by an AI-powered system are displayed at a branch of Lotte Mart in Seoul in July. [YONHAP]

Can AI help retailers deliver the perfectly firm peach or the sweetest melon? Domestic marketplaces think so.

Supermarket Lotte Mart and its sister chain of smaller stores, Lotte Super, have been sorting peaches and nectarines through an AI-powered system that utilizes deep learning algorithms to more accurately measure the sweetness and texture of the fruit as well as check for any defects. The move is part of a promotional campaign launched on July 18 that promises peaches and nectarines of even quality.

Last year, the two marketplaces first implemented a device that checks the sweetness of peaches without needing to harm their skin or flesh, and bumped up the standard weight of a peach eligible to be sold on the marketplace — specifically, the peach varieties harvested at a medium level of maturity, usually in mid-July — by 10 percent to reach around 250 grams (9 ounces).

This year, the AI system with the fruit sugar meter has been updated with deep learning algorithms that constantly obtain and compare data to measure the quality of produce, according to the two retailers' operator, Lotte Shopping. Two cameras take three to four photos of the peaches per second as they move on a conveyor belt to detect defective products. The data set of anomalies that deviate from those of a standard peach is constantly fed through the model. Images of damaged products are analyzed in tandem with data obtained by the fruit sugar meter to improve the accuracy and speed of sorting.

Peach with a split pit [DAEGU GYEONGBUK NEUNGGEUM NONGHYUP]

The AI system checks the peaches' potential defects in terms of color, size, sweetness and damages with precision, the retailer said. Not only can it identify fruits sporting unusual shapes, bug bites or mutations, but it can also sort out peaches or nectarines whose pits have split before fully maturing, which are hard to detect with the naked eye. The split pit creates a hollow center inside or around the seed, making the stone fruit unmarketable as well as potentially creating mold within the flesh.

The system is able to check for split pits; the cameras and sensors capture multiple images of the peaches from various angles and search for cracks near the stem, which are indicators of the defect.

Lotte Mart and Lotte Super sold nectarines that had been AI-sorted by weight, color and quality through its Re:Fresh campaign last July. This year, the system will also sort a large range of crunchy peaches including Changbang, Daewol and Gyeongbong varieties. Upgraded technology will be implemented to sort softer varieties in the future.

Lotte Shopping’s “Re:Fresh” project, launched in 2022, is the conglomerate's attempt to differentiate its produce from that of competitors by emphasizing its freshness. The project involves ensuring high quality for agricultural, fishery and meat products by tracking and improving on all stages of logistics, according to the group — from the place of production to the markets where customers make their choices. Lotte Shopping is also running campaigns for oriental melon, watermelon and abalone this year.

Oriental melon are displayed at a supermarket in Seoul in May. [YONHAP]

Supermarket chain Homeplus also sells oriental melon and tangerines that were sorted through an AI-powered system, which was implemented for the melons in February 2023 and the tangerines in December. The retailer also uses a deep learning algorithm that analyzes images taken of the fruit from multiple camera lenses to more accurately check for quality.

“The AI system checks the weight, sweetness, water content and ripeness of the fruit,” a Homeplus spokesperson said.

BY KIM JAI-HAK, KIM JU-YEON [kim.juyeon2@joongang.co.kr]