Sign in Subscribe
Periodical publishing

Robot-Aided Quality Inspection of Plastic Injection Molding Parts

Strategic Insights

— 1 min read

Researchers at the University of Applied Sciences in Cologne, Germany, have developed a new robotics system that can detect defects in plastic injection molding parts with high accuracy. The system combines an automated inspection procedure with an anomaly detection approach using artificial intelligence. The researchers used a 7-DoF robotic manipulator to automate part handling in front of an industrial optical camera sensor, capturing images to train a PaDiM anomaly detection network to reconstruct normal images of the parts. The results showed that the system can detect various defects with up to 100% defect detection rates while maintaining approximately 91% specificity using a small dataset of 117 parts.

Key Takeaways:

  • Researchers at the University of Applied Sciences in Cologne, Germany, have developed a robotics system that combines automated inspection and anomaly detection to inspect plastic injection molding parts.
  • The system uses a 7-DoF robotic manipulator to automate part handling and a PaDiM anomaly detection network to reconstruct normal images of the parts.
  • The results showed that the system can detect various defects with up to 100% defect detection rates and maintain approximately 91% specificity using a small dataset of 117 parts.
  • The research was conducted by Nicolas Kaulen, Dario Luipers, Laurenz Strothmann, and Anja Richert from the Faculty of Process Engineering, Energy and Mechanical Systems, Institute of Product Development and Engineering Design, University of Applied Sciences Cologne.
  • The study demonstrates the potential of robotics and machine learning in improving quality inspection in industrial environments.
  • The system can be used in various industries where plastic injection molding parts are produced, including automotive, aerospace, and electronics.

Statistics:

  • 100% defect detection rate for various defects
  • 91% specificity using a small dataset of 117 parts
  • 7-DoF robotic manipulator used for automation of part handling
  • Technical specifications of the PaDiM anomaly detection network and its reconstruction capabilities of normal images of parts

Sources:

  • Journal of Engineering. 2025; 2326.
  • Robot-Aided Quality Inspection of Plastic Injection Molding Parts Using an AI Anomaly Detection Approach in an Industrial Environment. Engineering Proceedings, 2025, 82(1):115. MDPI AG.
  • https://doi-org.sdpl.idm.oclc.org/10.3390/ecsa-11-22207.
  • University of Applied Sciences, Cologne, Germany. Faculty of Process Engineering, Energy and Mechanical Systems, Institute of Product Development and Engineering Design.

Read more