Consulting

Enhancing Upstream Inspection with AI

The Critical Role of Upstream Inspection in the Capping Process


Once you know what normal is, detecting the abnormal is easy.. Even with the toughest products. our anomaly detection platform for visual inspection Advisum AI, powered by boon AVIS, trains on defect-free product instead of getting lost in unknown defect types and other anomalies.

Current Limitations in Capping Machine Inspection

Traditional systems for capper inspection rely on rules-based camera setups that require detailed programming for each defect, often necessitating vision experts and OEM support for recipe training. These systems also face issues with outdated software and hardware, limiting their ability to adapt to evolving production needs. These limitations lead to inefficiencies in production, missed defects, and increased costs due to rework or recalls.

Advantages of Using AI for Upstream Inspection

To address these challenges, Genesis Packaging Technologies and Boon Logic developed Advisum AI, powered by Boon Logic’s AVIS platform. This solution uses unsupervised machine learning to learn from normal, compliant products and detect defects with a high degree of accuracy. Advisum AI can be retrofitted to existing capping machines with rules-based camera systems or deployed as a standalone inspection station for machines without cameras.

Advisum AI Offers Several Advantages Over Traditional Inspection Systems
  1. Quick and Efficient Recipe Training
    Advisum AI uses compliant units for training, learning from normal product variations without requiring extensive libraries of pre-labeled defect images. Training is completed in minutes, significantly reducing setup time.
  2. Optimized Defect Detection with Reduced False Rejects:
    Using unsupervised machine learning, Advisum AI accurately detects defects such as misaligned caps, missing stoppers, and cracks while minimizing false rejects.
  3. Simplified Validation and Compliance:
    Once the machine learning model is trained, it can be locked for straightforward qualification and validation.

Advisum-AI can detect a range of defects during both pre- and post-capping inspections, including:

  1. Misaligned or Cocked Caps
    Detects caps that are improperly seated, which can compromise container closure integrity.
  2. Missing or Raised Stoppers
    Identifies vials with missing or improperly inserted stoppers, preventing contamination or leakage.
  3. Cracked or Chipped Vial Necks
    Recognizes damage to vial necks, such as cracks or chips, caused during the capping process.
  4. Particulate Contamination
    Detects foreign particles, such as metal or plastic shavings from the capping machine, trapped under the cap.
  5. Deformed Cap Skirts
    Identifies defects in the cap’s skirt, such as dents or irregularities, that can impact sealing and vial appearance.
  6. Discoloration in Product or Wrong Cap Color
    Can correctly identify the proper appearance of your product and also determine if the wrong color cap has been applied.

Advisum AI is flexible and can be integrated into different types of capping machines, allowing manufacturers to upgrade inspection capabilities without costly overhauls or downtime:

  1. Retrofit Capping Machines with Existing Camera Systems
    Advisum AI can be integrated with existing hardware, maximizing the return on investment.
  2. Retrofit Downstream Semi-Automated Visual Inspection Systems
    Replace human inspectors by automating semi-automated visual inspection (SAVI) machines.
  3. Standalone Inspection Station
    Advisum AI can also be installed as a standalone inspection solution for machines without built-in cameras.

Proven Results: Case Studies and Field Performance

Testing conducted by Genesis Packaging Technologies and Boon Logic demonstrated that Advisum AI reliably detects defects on capping lines. In one case study involving a vial capping line, Advisum AI identified over 90% of defects while maintaining a false reject rate below 1%. It effectively detected issues such as crooked caps and missing stoppers, while significantly reducing false rejections, which previously led to production delays and increased costs.


Replacing Human Inspection of Powder Filled Molded Glass with AI Based Inspection

A CMO recently failed numerous AQL's from semi-automated visual inspection and was looking for a solution to inspect their powder-filled molded glass container. Learn how the Advisum AI visual inspection system demonstrated a detection accuracy of over 98% and a 3% false reject rate and significantly outperformed human inspectors solving their issue.


Conclusion

Advisum AI helps manufacturers identify defects earlier in the production process, preventing costly product waste, reducing rework, and avoiding production delays. By enabling real-time adjustments at the capping stage, it ensures higher product quality and process efficiency, while minimizing the risk of defects reaching final inspection or distribution. This proactive approach to inspection not only safeguards product integrity but also enhances overall operational efficiency in pharmaceutical manufacturing.