Deep Inspection Collatio Features
Optimized for high-mix low-volume manufacturing
Collatio doesn’t require much good / defective product data to be put into use, so it can be introduced to sites where the number of production lots is low and the specifications change a little every time.
After installation, a variety of products can be developed smoothly with minimal labor.
One piece of good product data can be used by applying learning from other products. If there is an error, the AI will learn each time, constantly improving operational accuracy.
Can be installed into existing lines
Our product is not a package, it is a system that makes use of the customer’s current production line. With the capacity to design for a variety of different work sites, we offer products with long-term operation and upgrades in mind.
Three Steps of Comparative Inspection
-
STEP.1 Creating pair images
Creating pair images of good quality and bad quality versions of the same parts in the same area.
-
STEP.2 Show defective areas
Using Rist’s original comparative neural network, the AI learns the characteristics of good and bad areas.
-
STEP.3 From verification to installation
Performing verifications and introducing it to the production line while actually running the systems. If provided with images, the AI can still detect OK and not OK examples of products it hasn’t previously learned.
* Images are for illustration purposes.
case.01Abnormality detection in printed materials using comparative inspection
The AI can learn to recognize subtle color and shade differences due to printers and scanners, and printing abnormalities such as blurring, dirt, and scratches, identifying only abnormalities that need to be detected.
