VISION – automatic defect detection system

Automatic defect detection based on photos. With the help of a data-driven algorithm, the system quickly learns to recognize and localize various defects. These can then be reworked automatically.

GENERAL FUNCTIONALITY

SCIIL VISION is able to detect teached defects, such as wrinkles, scratches, staining and creases automatically from photos.

Detected defects can not only be saved and analyzed but also send to automatic rework stations like, in our example, a steaming robot.
Here you can find the product video on Youtube. 

TEACHING PROCESS

  • Data-driven algorithm that transfers ‘expert-knowledge’ to artificial models
  • Prepared data fed to the neural network during its training
  • Approx. 20-25 good and bad examples needed for training

DEFECTS DETECTION AND SEGMENTATION

  • Detection with region-based approach
  • Final prediction with separate model

ANALYSIS AND REPORTS

Defects are saved in the database for future analysis from remote clients.

Traceability data, statistical information, data grouping and reports with export functionality.

ADVANTAGES

Most existing systems compare current photos with hundreds or thousands of predetermined example pictures. This causes immense amounts of data and increases the cycle times.

SCIIL VISION uses artificial intelligence and learns. That needs less time and less data.

Due to different lighting and perspectives nearly 100% of the defects can be found.

REWORK

The data collected during the analysis can be used for automatic rework processes.

The system knows the exact region and failure and and can thus control the corresponding tool.

After the rework a second picture is taken for control.

Kontakt

SCIIL AG
Marktstraße 83
56564 Neuwied
+49 (0) 2631 999 88-0
+49 (0) 2631 999 88-99
sales@sciil.com

Request to VISION


    I agree that my details from the contact form will be collected and processed to answer my request. The data is stored for 10 years or deleted on request.
    Note: You can revoke your consent at any time for the future via e-mail to datenschutz@sciil.com. Detailed information on handling user data can be found in our privacy policy.