AI-Driven Food Sorting Robot for Peach Processing
Robovision collaborated with integration specialist QING to develop a peach sorting robot with machine vision, successfully creating a fully automated quality control system with a level of precision and speed beyond human capability.
Robovision's Indurion Suite allows us to create, implement and maintain AI models within food production in an accessible way. This provides opportunities for food production companies to use the application themselves without relying on outside expertise.
Teun Keusters Product Manager, QING Food Automation
Automated quality control
Agrophoenix, based in Greece, is a leading company in the fruit processing sector, specializing in processing peaches into consumer products. The company approached engineering company QING because they needed a solution to optimize product quality, minimize food waste, and automate quality control to address labor shortage. Automation would also increase efficiency and reduce production and personnel costs.
Integrating the Robovision AI platform
QING partnered with Robovision and developed a Vision AI application using the Robovision’s Indurion Suite to automate the peach sorting process. They tested the integration of various units, including lighting with cameras, robotic arms and conveyors, and installed the control software supporting our AI platform.
Qualify and sort peaches
Setting up and training the Vision AI algorithms was partially performed at QING in the development lab. Then, the system was shipped, rebuilt, and tested at the end customer's site. The application was trained to qualify and sort the peaches for abnormalities, such as remaining peach kernels or a crushed peach.
Meeting high quality standards
Agrophoenix’s expansion illustrates a remarkably rapid transition from an initial pilot setup to a large-scale, automated production infrastructure. Following the successful launch of a single AI-driven sorting line in May 2024, the Greek fruit processor scaled up operations after ten months of seamless performance by commissioning seven additional lines in June 2025.
This expansion ultimately extended across three sites, featuring a total of 12 robotic arms and a combined processing capacity of over 500,000 peaches per hour. Through this accelerated automation, Agrophoenix not only ensured consistent product quality and high sorting accuracy but also reduced its reliance on hard-to-find seasonal labor by the equivalent of 40 full-time positions (FTEs).
Inspection and sorting applications
QING has developed the advanced STAQ Automation Framework- 'See, Think, Act'- specifically designed for inspection and sorting applications. Vision AI technology, powered by the Robovision Indurion Suite, enables the 'Think' step of this framework. It optimizes and automates the entire sorting process. For the project at Agrophoenix, extensive testing has been conducted on the integration of various components, such as precision lighting, high-resolution cameras, robotic arms, and conveyor and tracking systems. Additionally, the control software has been implemented, providing strong support for the AI-driven platform and ensuring smooth, efficient operational workflows.
Maintain AI models within food production
QING enables food producers to apply vision AI in an accessible way, integrating Robovision’s Indurion Suite to easily train, implement, and maintain AI models within food production. QING's approach is centered around collaboration: they work together on a solution and provide support in using the Vision AI platform. This allows companies to gain knowledge and experience with AI, with the right guidance, after which they can adopt the technology at their own pace and maintain control.