Artificial intelligence improves cherry sorting: more quality and less waste

31 Mar 2025
496

The sorting and packaging sector of cherries is undergoing an unprecedented transformation thanks to artificial intelligence (AI). From advanced sorting systems to optimized quality control and predictive maintenance, AI is making packaging lines faster, more precise, and more efficient.

This change is particularly relevant in high-speed and high-volume processing operations, where precision is crucial to maximize yields and reduce waste.

The Challenges of Sorting

Cherries are naturally diverse: size, color, shape, and defects represent a complex set of variables to analyze. Additionally, their delicate nature makes them vulnerable to damage such as cracks, rot, bruises, wind scars, sunburn, and stem deterioration.

Optimal management of these aspects is crucial to ensure that only the best fruits reach the market. Over the past two decades, advancements in vision systems for sorting have already improved evaluation capabilities, but today AI is taking quality assessment to the next level.

Artificial Intelligence in Service of Quality

AI is not an entirely new concept, but the recent development of deep learning has made this technology more accessible and effective. GP Graders, a leading company in the sector, has developed radiai, an innovative system based on convolutional neural networks (CNN) that is revolutionizing cherry sorting.

"We have cataloged and annotated hundreds of thousands of cherry defect images over the past five years," explains Stuart Payne, director of GP Graders. "Thanks to this vast database, radiai can identify and classify defects with extraordinary accuracy, surpassing traditional rule-based systems."

Radiai's deep learning system not only recognizes imperfections such as discoloration or cracks, but it can also adapt and improve over time, continuously refining its recognition capabilities.

Concrete Impacts on the Industry

The results of radiai technology are already tangible. Companies using gpVision with radiai report near-perfect detection of stems, eliminating measurement errors that previously compromised fruit classification.

This has led to improved diameter precision and increased packaging efficiency. Another groundbreaking aspect is radiai's ability to distinguish between real defects and visual artifacts.

For example, dry or darkened wood on a stem, which in the past could have been mistaken for a crack and caused the rejection of a perfectly good cherry, is now accurately recognized, significantly reducing waste.

One of the most significant advancements is the detection of fresh microfractures, which are difficult to identify with traditional systems. Thanks to an ever-expanding database, radiai has nearly eliminated this issue, allowing producers to ensure superior quality.

Automation and Simplification for Operators

In addition to improving fruit sorting, AI is also simplifying the work of operators. In the past, the classification process required complex manual adjustments. Now, with gpVision and radiai, operators can control the system through an intuitive interface with just a few simple commands.

This reduces training time and enhances productivity. "Today, achieving accurate sorting is easier than ever," emphasizes Payne. "Operators only need to review the data, make minimal adjustments using intuitive controls, and let the system handle the rest."

This not only speeds up the setup process but also ensures consistent quality and unprecedented efficiency."

The Future of Cherry Sorting

While the world debates the impact of AI on everyday life, in the fresh produce industry, this technology is already making a difference. The processing of over 10,000 tons of cherries last season has demonstrated that AI can significantly improve the final product's quality, reduce waste, and increase consumer satisfaction.

With such promising results, the future of sorting cherries seems increasingly tied to artificial intelligence. And for companies in the sector, investing in this technology could be the key to staying competitive in an increasingly demanding market.

Source of text and images: FreshPlaza.com


Cherry Times - All rights reserved

What to read next

Early cherries in Argentina: a study focuses on dormancy to anticipate the harvest

Varieties

21 Oct 2024

The study focused on an experimental block of 2,000 early cherry plants, including traditional cultivars such as Brooks and Lapins. The Royal Lee variety from the Zaiger program was the first to be ready for harvest on September 24, with soluble solids reaching up to 21° Brix.

Optimising nitrogen management in cherry trees: new approaches to maximise yield according to the University of California

Tech management Press review

25 Jan 2024

The project results promise new standards for nitrogen management in the cherry industry. Fertiliser use can be optimised and N losses reduced if applications follow tree demand and seasonal uptake patterns.

In evidenza

Cherry standards reform stalled in US: uncertainty grows for growers

Quality

28 Jul 2025

The Washington Cherry Marketing Committee proposes updates to cherry sizing and packaging standards, but USDA approval is stalled. Political pressure and regulatory gridlock hinder innovation, crop insurance protection, and the competitiveness of the cherry sector.

German cherries as superheroes: sustainability meets mindful consumption

Consumption

28 Jul 2025

On July 16, Germany honors cherries as a symbol of sustainability and local pride. With 45,000 tons harvested in 2024, this fruit faces climate and global challenges but relies on local supply chains, innovation, and mindful consumption.

Tag Popolari