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
