How AI and Hyperspectral Imaging Helped Cut Food Complaints by 90%

In the field of food quality and safety inspection, the stakes are high. Food producers are obligated to manage the quality and safety of their products, which poses the demand for accurate foreign material and contaminant detection in manufacturing processes. The failure to reliably detect quality issues in food, such the presence of foreign materials, can cause considerable reputation damage, financial losses, and risks to consumer safety.

ELROILAB is a South Korean deep-tech company founded in 2020 that develops AI-driven hyperspectral imaging (HSI) solutions for industrial applications to help manufacturers meet this challenge. The company has created the world’s first proprietary AI deep learning software dedicated to hyperspectral imaging, enabling automatic detection of foreign materials in food and other manufacturing processes with Specim’s advanced HSI technology. Today, major Korean food companies rely on ELROILAB’s solutions to ensure the quality and safety of their products.

Bones in Minced Chicken Threaten Product Quality and Consumer Trust

One of ELROILAB’s customer projects focused on improving the detection of tiny foreign materials in baby food. A major Korean producer faced challenges on a production line using minced chicken. The company was receiving significant customer complaints about chicken bones found in the product, which posed a serious consumer safety risks and damaged brand trust and reputation.

The customer needed a solution that could reliably detect chicken bones and other foreign materials in minced chicken. White bones are nearly indistinguishable from the surrounding meat, making them impossible to spot with the naked eye or standard RGB cameras. In addition, because food safety was at stake, the solution had to detect not only bones but also a wide range of potential contaminants. As a result, the company depended on manual inspection by touch—a slow, labour-intensive, and inconsistent process that left room for errors. ELROILAB addressed the challenge with a hyperspectral imaging (HSI) and AI-based solution, overcoming the limits of conventional inspection methods.

Hyperspectral Imaging + AI Catches What Eyes Miss

The objective of the project was to automate and enhance the accuracy of detecting extremely small foreign materials, which had proven difficult using conventional visual imaging and manual inspection. To address this need, ELROILAB presented their client with a solution that combines the use of Specim’s HSI cameras with ELROILAB’s own innovative deep learning-based AI analytical software.

In-line Hyperspectral Inspection of the Entire Product Stream

Specim’s HSI cameras use spectral information to distinguish and detect foreign materials in food with high precision. By analysing the unique spectral signatures of different substances, they replace visual and manual inspections with non-invasive, real-time monitoring of the entire product stream.

An HSI camera analyses a material’s chemical composition based on its unique reflectance spectrum, rather than relying on visual appearance. This allows it to identify foreign materials that are visually indistinguishable to the human eye or standard RGB cameras. In addition, HSI cameras enable the real-time analysis of the quality uniformity of raw materials and tracking changes in composition.

ELROILAB uses Specim’s industrial hyperspectral cameras, the Specim FX10 and Specim FX17, together with its proprietary AI software, SPECTRAL AI in handling complex hyperspectral data. The software is focused on utilising huge data of hyperspectral images. It adopts a deep learning framework optimised for detecting foreign materials in food, continuously learning from data of irregular foreign objects that are frequently detected.

Once foreign materials are identified, they are removed using one of three extraction methods:

• Dropdown method – isolates and drops only the foreign material/contaminant while allowing the main product to continue flowing.
• Air-blow method – removes foreign material/contaminant with a precise burst of air.
• Delta robot method – detects and vacuums only the foreign material/contaminant.

Impact: 90% Fewer Quality Issues

By applying Specim’s HSI solution, ELROILAB was able to accurately analyse the spectral differences between food raw materials and foreign substances across various wavelength ranges, greatly enhancing the reliability and efficiency of foreign material detection.

With hyperspectral cameras, chicken bones and other foreign materials that had been extremely difficult to detect using conventional methods in minced chicken meat could now be identified with ease.

The impact for the client was substantial: ELROILAB’s hyperspectral solution enabled a significant reduction in manual inspection staff, while customer complaints decreased by nearly 90%.

Using HSI technology in this application delivered several concrete benefits:

• Nearly 90% reduction in customer complaints
• Minimised food safety risks
• Improved accuracy of foreign material detection
• Reduced labor costs through process automation
• Ensured consistent quality standards
• Enabled production data accumulation and process optimisation

Why Specim: Proven HSI Technology, Industrial Expertise, Flexible Integration

ELROILAB chose Specim as their partner for its quality inspection system, relying on Specim’s globally proven HSI technology and extensive industrial experience. Specim’s consistent performance across diverse wavelength ranges and flexible system integration capabilities made it the clear choice.

“As the global leader in hyperspectral imaging solutions, Specim products deliver high-performance hyperspectral data that can clearly distinguish materials with similar colors or those difficult to identify with the naked eye. They are a highly reliable choice for fields requiring precise quality analysis and detection.”

–Jay Lee, Marketing Manager, ELROILAB

What’s Next: Scaling to Recycling & Manufacturing

The future looks bright for the combined use of Specim’s HSI technology with ELROILAB’s AI analytical capabilities, extending beyond food safety. ELROILAB plans to expand the application of its hyperspectral AI solutions to various industry applications, including recycling and manufacturing.

In close collaboration with Specim, ELROILAB aims to enhance their competitiveness in the global market and provide greater value to their customers through innovative inspection automation and quality management solutions.

Conclusion

In this project, ELROILAB needed to achieve high reliability and accuracy in a demanding food inspection environment, and Specim’s HSI solutions met the challenge. The project showcases the power of HSI for the detection of foreign materials in food manufacturing processes yet scratches only the surface of what’s possible with this technology.

---

---