90% accuracy in detecting all critical defects
Kompass (AI) and Tigris (Modular Vision Infrastructure)
0.01% false positives, leading to near-zero wastage
Overview
Britannia is one of India’s oldest and most renowned food companies, specializing in the production of biscuits and other bakery items. With headquarters in Kolkata, the company distributes products nationwide and exports globally.
As biscuit demand soared, Britannia needed a faster, more reliable solution to reduce manual inspection fatigue and catch critical defects before they could affect customers—or the company’s bottom line.
The Opportunity
For Britannia, Manual defect detection was no longer efficient or fully reliable:
- Over 12,000 biscuits produced per minute, requiring 24/7 inspection.
- Operators found it challenging to spot cracks, chipped edges, or burn marks consistently at such high speeds.
- Even minor errors passing downstream could lead to returns, complaints, and damage to brand reputation.
- The biscuit inspection process was entirely dependent on human labor, increasing the risk of fatigue-driven mistakes.
Jidoka’s Approach
Jidoka installed Kompass and Tigris to automate and optimize Britannia’s inspection workflow:
- Vision-Based AI Inspection:
Three cameras spanned 27 columns of biscuits on the conveyor. Each camera captured over 400 images per minute, sending them to Jidoka’s AI engine for real-time analysis.
- Defect Identification & Classification:
Cracks, chips, and surface deformities were quickly flagged, enabling precise ejection of faulty biscuits.
Visual AI-based biscuit inspection at Mondelez facility (representative image). Britannia uses a similar high-speed inspection system from Jidoka
Jidoka’s hardware, Tigris, connected to Britannia’s existing lines with minimal disruption. Tigris’ automatic rejection unit removed defective pieces on the fly.
Built to handle diverse biscuit SKUs and changing process variations, the solution has now rolled out across 50 lines in India.