MLAI Digital Pvt. Ltd.

AI Powered Video Anomaly Detection
- Manufacturing Industry

Data Analytics using Speech Recognition & NLP

1. Quality Control:
a. on Plant floor, VAD is used to monitor product quality and detect defects in real-time
b. Real time anomaly detection helps enhance quality standards, with near-zero human intervention.

2. Operational Efficiency:
a. Help identify inefficiencies, bottlenecks, or deviations from standard design / operating procedures
b. Accordingly , corrective actions can optimize processes & reduce downtime 
c. Reduced Production time

3. Predictive Maintenance:
a. VAD helps monitor equipment & infra for signs of wear, damage, or malfunction, allowing proactive maintenance 
b. Reduces breakdowns and downtim

Project 1 : Cylinder Head Inspection

Manual Process Earlier

  • A cylinder head with 12 holes comes on assembly line
  • Labour has to inspect if all 12 holes has 12 rings properly put in holes
  • If any ring in any hole is found wrong, entire cylinder head is rejected

AI Video Anomaly Detection

  • After Automation, the image / Video stream is fed to AI tool
  • Tool now tells if the ring is put right ( green ) / put wrong ( red ) or ring is missing ( Blue )
  • AI model is trained to identify anomaly and it improves with time, by self-learning

Project 2 : Water Pump Inspection

Manual Process Earlier

  • L shaped water pipe is manually checked for blockage by throwing light from one end and checking light from another.

  • If light comes from other end, Pipe is ok. Else it is not ok ( blocked ) or partial ok

AI Video Anomaly Detection

  • Image / Video stream is fed to AI tool and model gives out as OK / Partial / Not OK

  • Camera is set on conveyor belt and stream Is fed to tool. No human intervention in quality check now