its headquarters in Mumbai, Marico is present in over 25 countries across Asia and Africa. Marico touches the lives of 1 out of every 3 Indians, through its portfolio of brands such as Parachute, Saffola, Saffola FITTIFY Gourmet, Saffola ImmuniVeda, Saffola Mealmaker, Hair & Care, Parachute Advansed, Nihar Naturals, Mediker, Pure Sense, Coco Soul, Revive, Set Wet, Livon, Just Herbs, True Elements and Beardo. Marico has 8 factories in India located at Puducherry, Perundurai, Kanjikode, Jalgaon, Paldhi, Dehradun, Baddi, and Paontal Sahib.
Client Challenges and the Business Need
Being a multi-national organization selling hair oil products, Parachute brand, in a competitive market, it was important to maintain high-quality packaging. This included no printing defects on the Parachute hair oil bottles during the packaging process. The brand’s reputation and customer satisfaction also depended on accurate printing on the bottles.
Advertisement
- The existing process was manually driven. Due to manual inspection for defect detection of the print on the bottles, the Client was facing a very high percentage of rejections shipped to consumers. This was due to human error and fatigue experienced by QC personnel as there was the need for round-the-clock inspection with the production of close to 1.5 lakh bottles per day (at 120 bpm).
- Unreliable detection rate on branding label defects that included misalignment labels and smudge printing gave rise to other challenges. These issues led to returns from the market for the perception of counterfeit products and a decline in the appeal, which eventually led to revenue losses for the Client.
- The manual QC process was time-consuming and led to production bottlenecks and a significant reduction in throughput.
Advertisement
To address these challenges the Client wanted to implement an AI-driven automated quality inspection system which enabled achieving near-100% accuracy in defect detection of printing on the bottles. The Client partnered with Jidoka Technologies, which is among the pioneers in the field of automated cognitive inspection for manufacturing to address the need first at its Pondicherry plant.
Solution
Jidoka Technologies’ team of Engineers had discussions with the Client’s Packaging team to understand the requirement before configuring the most appropriate solution. The team then conceptualized a solution that comprised the below components,
- Software: Jidoka leveraged its cutting-edge software platform Kompass with Deep Learning AI for quality inspection of printing for 360 degrees of the bottles.
- Camera Set-up: Infrastructure with two high-resolution cameras was strategically installed to capture 360 degrees images of the bottles in near real-time as they moved through the packaging line. The images captured from the cameras were stitched and sent to Jidoka Technologies’ AI/ML engine for identifying possible defects and the result was made available within 0.8 seconds.
- Automated Rejection Process: The output generated by the system is sent to an automatic rejection system which removed the defective bottles from the production line. An alert mechanism was triggered to indicate the continuous occurrence of any particular type of defect.
Business Benefits
With the implementation of the AI-enabled automated quality inspection system, the Client’s packaging plant at Pondicherry experienced significant improvements.
- 99.7% accuracy in detecting all defects.
- 4% improvement in defect detection compared to manual inspection.
- USD 25,000 cost savings per year with process improvements in addition to minimizing product returns.
- Enabling immediate process corrections to reduce defects occurrence through the “ANDON” system.
- Improved compliance with printing regulations and highest quality standards to eliminate consumer reservations about counterfeit products.
- Client wants to deploy the solution at its Perundurai Packaging Plant as well.
Please visit, https://jidoka-tech.ai/
