Nichirei Foods - Automating Ingredient Inspection for Frozen Foods
Using Optimal Multi-Spectral Cameras with SWIR Image Sensors

Application: Food inspection
Company name: Nichirei Foods Inc.

Home
Case Study
Nichirei Foods - Automating Ingredient Inspection for Frozen Foods Using Optimal Multi-Spectral Cameras with SWIR Image Sensors

Public perception of frozen foods is shifting significantly, evolving from a convenience option to providing greater ingredient breadth and variety to daily diets. Amid rising concerns about the soaring cost of living and the stability of the food supply, however, the food industry is under pressure to meet increasing demands while also ensuring high product quality and production efficiency. With food safety being the highest priority, there is also the critical issue of enhancing quality control systems to reduce the risk of foreign matter contamination.
These factors, particularly are escalating expectations for more accurate and reproduceable inspection methods, particularly alternatives to conventional visual inspection and general image inspection.

Reference image depicting frozen food production

About Nichirei Foods

Nichirei Foods is a leading company in frozen foods and other food products, serving both consumers and commercial customers. The organization boasts a long history of freezing and processing technologies, offering safe and trustworthy products that contribute to more varied and enjoyable diets.

Challenges: Tendency to overlook foreign matter during ingredient acceptance inspection

Nichirei Foods has been working to advance the automation of food production processes, developing and introducing inspection equipment that uses image processing technology and AI to detect objects that are difficult or impossible for the human eye to perceive. Its acceptance inspection of ingredients for frozen foods involves both human visual inspections and camera-based image inspections, but the company identified challenges such as:

- Difficulty detecting foreign matter, particularly transparent materials like plastic shards or materials that are similar in color to the ingredients and difficult to distinguish
- Inspector fatigue which reduces concentration after long working hours, introducing the risk that sub-quality ingredients may be overlooked

To maintain consistent inspection performance, the company needed an additional inspection method, one that could automatically and reliably detect foreign objects that are hard to discern with the human eye or conventional color cameras.

Solution: Optimal Multi-Spectral Cameras with Multi-Band Filters

Sony Semiconductor Solutions Corporation (SSS) proposed the concept of an optimal multi-spectral camera that combines a SWIR image sensor and a multi-band filter as a solution for these issues.
Link: About Optimal Multi-Spectral Cameras

SWIR image sensors are sensors that capture light in the visible to short-wavelength infrared (SWIR) band of 0.4 μm to 1.7 μm.
Photography with a multi-band filter can capture four different wavelengths of information. Incorporating information from multiple wavelengths makes it possible to perform high-precision detection - including discerning foreign matter that closely resembles the ingredients in color or shape - and to visualize the moisture content of substances.

Nichirei Foods has been using these optimal multi-spectral cameras and its own proprietary image processing technology to achieve more advanced foreign matter detection during the ingredient acceptance inspection process.

Comments from a representative of Nichirei Foods
Yusuke Ago, Machinery Development Group, Engineering Strategy Department

We adopted SSS’s optimal multi-spectral camera concept because we appreciated how it utilizes multiple short-wavelength infrared (SWIR) to achieve something that was difficult or impossible with conventional X-ray or visual inspections: distinguish between food and resin.
In actual verification, it has achieved higher accuracy in detecting transparent resins, which has led to higher levels of quality assurance. What’s more, by combining AI and image processing, it doesn’t just detect foreign matter; it also helps to ensure more accurate decision-making and greater stability in actual operation.
Introducing this system has allowed us to take a new approach to detection and solve issues that have plagued conventional technologies.
I believe that multi-spectral imaging is very valuable as a foundational technology to support safety and reliability in food production. The system based on this technology will be rolled out to our other plants going forward and we will strive to achieve even higher quality across the company.

This demonstrates why inspection using optimal multi-spectral cameras is highly rated for its ability to make judgments based on differences in material, regardless of similarities in color or shape.

Diagram explaining the concept of optimal multi-spectral cameras

Images captured for foreign matter detection assuming acceptance inspection of green onions

Explanation of foreign object inspection of green onions during the acceptance inspection process, using a standard color camera

Explanation of foreign object inspection of green onions during the acceptance inspection process, using an optimal multi-spectral camera

Results: Achieve both higher-accuracy inspection and labor saving

By incorporating optimal multi-spectral cameras into image inspection, Nichirei Foods has successfully used automation to reduce the burden on its employees.

It has also produced the results below:

- Improved the accuracy of foreign matter detection during ingredient acceptance inspection
- Stabilized the inspection process by increasing speed and reducing variability
- Achieved labor-saving on the inspection line and reduced the burden on inspectors

The ability to conduct high-accuracy foreign matter inspection from the ingredients stage has helped Nichirei Foods to stabilize its entire manufacturing process, resulting in stronger systems for providing consumers with frozen food products that are safe and trustworthy.
This system is currently in operation at its Funabashi plant near Tokyo, and the company expects to roll it out to more production lines going forward.

Download

A List of Cameras with SWIR Image Sensors

Download a list of cameras that incorporate Sony's SWIR image sensors. The optimal multi-spectral cameras are described in the Notes column as Note: Based on Sony's "Optimal Multi-Spectral Camera" concept.

Imaging Setup and Multi-Band Image Processing Techniques for Optimal Multi-Spectral Cameras

For actual verification of optimal multi-spectral cameras, key points of photography and image processing techniques are introduced, along with examples of photography. Sample images of actual shots are also available for download.

Product & Service Inquiries

Click here to request for a datasheet/quotation (RFQ)!

* This button will redirect you to the salesforce.com Co., Ltd. website,
which we have entrusted.

E-mail Newsletters

Find the latest information on our newsletter for industrial and security applications.

Subscribe Now

Nichirei Foods - Automating Ingredient Inspection for Frozen Foods Using Optimal Multi-Spectral Cameras with SWIR Image Sensors