INDUSTRY

F&B

Laon People's solutions improve productivity by quickly, accurately, and transparently identifying defects.

Strict hygiene standards are required in the food and beverage industry. In addition, flexible and safe process operations are also important. LAON PEOPLE's solution can be applied to various fields such as counting quantity and distinguishing three-dimensional objects. In particular, the solution is robust to inspect for appearance defects (crushing, breakage, contamination).

Experience the case of other companies that introduced LAON PEOPLE's artificial intelligence software first.

INQUIRY
sales@laonpeople.com+82-1899-3058ONLINE INQUIRY
LIST

Product Specific Specifications

Information
SCROLL >>
Not registered post.
Inspection for Foreign Substances in Beverages Information
Applied Solutions
Mars + LPMVC-EE120C
The beverage production line needs to be inspected for foreign substances inside the PET bottle before shipment. However, there are various difficulties in the examination stage. Bubbles and water droplets exist due to the characteristics of the liquid, which are often detected as foreign substances, or due to the characteristics of the PET bottle, light reflections occur and are mistaken. LAONPEOPLE's deep learning-based artificial intelligence vision inspection solution can increase efficiency and provide objective and accurate quality control by automating the previously visually inspected parts of workers
Accurate Detection of Foreign Substances and Robust Performance Against the Light Reflection and Bubbles
Most samples detected foreign substances well and showed robust performance against light reflection, blisters, and drops of water.
We tested 105 normal products and 95 defective products (a total of 200), and the results of the test are as follows.

1) Test for the Robust Detection of Foreign Substances in a Liquid
- 0 undetected cases
- Recognize both white and red foreign substances in the foreign image of 95 defective products.
- Solve the problem of difficulty in detecting foreign substances that are difficult to see even with a human eye as it is obscured by the contents by adjusting the threshold of foreign substance probability





2) Misrecognition Test for Bubbles and Water Droplets
- 2 over-detected cases
- 2 out of 105 images of normal products where bubbles and drops of water are recognized as foreign substances
- Solve the problem with adjusting the threshold of size of the foreign substances as the number and size of over-detected objects are very small





3) Misrecognition Test for Light Reflection
- 0 over-detection cases
- No misdetection of light reflection in all images of 105 normal products
- Show the robust performance at the light reflection





Inspection for the Defective Tofu Molding Information
Applied Solutions
NAVI AI - Mars
In fact, the largest food manufacturer in Korea, the client company, wanted to introduce a molding defect inspection solution to the tofu production line during production.

In addition to the different shapes and locations of tofu, the problem of defective molding was that noise (the shape of bubbles/packing materials) generated by image acquisition was difficult to implement as a solution of the existing rule-based method.

LAONPEOPLE provides a vision solution that enables faster and easier inspection through NAVI AI Mars, Deep Learning technology, which learns not only the areas of poor images but also the areas of quality.
Overview of Inspection
  • Inspection Field: F&B
  • Inspection Parts: Tofu
  • Inspection Type: Deformed/torn status
  • Configuration of Inspection Solution
    - Inspection Image: Set self-optical conditions and shoot
    - Inspection S/W: LAONPEOPLE NAVI AI Trainer, NAVI AI

 

Contents of Inspection
Defective Type of Metal Sheet
[Disformed Status] In case the shape of the product is different due to defective molding of tofu
[Torn Status] In case the tofu has been torn off during the production process





Features and Benefits of LAON PEOPLE
  • ​Fast inspection speed (up to 200fps based on 1.2m image)
  • Enable customized classification inspection for defect types as the inspector can participate in learning.
  • Be intuitive as the updating flow of program is conducted as if the actual person is inspecting.