Overview
Pernod Ricard, a global leader in premium wines and spirits, produces approximately150,000 bottles daily at its Gwalior facility. As demand grew, the company faced a critical challenge: ensuring flawless label quality on every single bottle. Traditional manual inspection, limited to unreliable sampling, risking product quality and the brand’s premium reputation.
Opportunity
Pernod Ricard’s manual defect detection process along the line was no longer reliable or sufficient to maintain its high standards.
- Over 150,000 bottles to be inspected daily
- Every text on the label needs verification against the approved template
- A single standardized inspection system is required for all label variants
- A delicate system capable of inspecting and rejecting bottles even during sudden halts or unexpected pile-ups
Even minor label defects could compromise brand image, mislead consumers, and risk regulatory compliance. Jidoka ensured flawless inspection—catching every defect with precision.
Jidoka's Approach
- Two-Station AI Vision Inspection System: Jidoka deployed a dual-station AI-powered vision inspection system to ensure end-to-end label quality verification.
- Station 1 – Label Verification Before Application:
Running labels from the printer are captured by high-resolution cameras before pasting. The images are processed through Jidoka’s KompassTM AI engine, which compares them against the master template for print accuracy. Optical Character Recognition (OCR) technology validates printed characters, ensuring legibility and compliance with regulatory text standards. - Station2 – On-Bottle Label Integrity Check: Two synchronized cameras capture the front and back surfaces of each bottle. Using advanced object detection algorithms, the system identifies defects such as barcode information mismatch, missing, wrinkled, or misaligned labels, guaranteeing consistent presentation across all bottles.
- Station 1 – Label Verification Before Application:
- Customizable User Interface (UI)
- A custom, Streamlit-based User Interface allows for rapid customization and deployment of new product templates.
- Designed for usability, it enables operators with minimal training to annotate inspection parameters and retrain AI models in real-time, ensuring swift change overs between different product lines and minimizing downtime.
- Integrated Control & Traceability:
The system integrates with MES controls and sensor-based automation to streamline operation.- Sensors continuously monitor the rejection station’s capacity and alert operators when nearing limits.
- A bottle traceability mechanism ensures that only defective bottles are diverted for rejection, maintaining accuracy and process integrity.
- The HMI allows quick adjustment of defect thresholds and operational parameters without interrupting production.
- Regulated Throughput & Safe Rejection:
- A bottle regulation mechanism prevents overflow and ensures optimal presentation to the vision system.
- The pneumatic finger rejection system safely and precisely removes defective bottles from the line without risk of damage or breakage, maintaining a smooth production flow.
Big Wins
With Jidoka's intelligent inspection infrastructure, Pernod Ricard realized significant wins.
300 Bottles/Minute
Guaranteed 100% reliable inspection at full production speed.
Universal Platform
A single, scalable solution for all label variants.
>30% Efficiency Gain
Dramatically higher accuracy than manual checks.
Intuitive UI
Enables simplified operation and rapid changeovers.
With Jidoka's Kompass-enabled vision AI system, Pernod Ricard eliminated label defects across its premium spirits portfolio, protected brand integrity, and ensured flawless product presentation. The deployment achieved near-zero defect escape, improved detection efficiency by over 30%, and set a new industry standard for automated quality inspection in high-speed bottling operations.






