99.5% ANPR Detection Accuracy at JSW Steel Dolvi Works
AI Video Analytics for High-Volume Vehicle Movement at a Steel Plant
JSW Steel Dolvi Works piloted Mikshi AI to automate vehicle identification, authorization checks, and PUC compliance using existing camera infrastructure.
- 99.5% ANPR detection accuracy
- 17,000+ vehicles monitored
- ~2,500 vehicles processed per day
- Authorized and unauthorized vehicle identification through database synchronization
- Invalid PUC vehicles flagged automatically
- Real-time licence plate display through an integrated LED display
About the Facility
JSW Steel Dolvi Works handles approximately 2,500 vehicles every day, creating a high-volume vehicle movement environment where automated identification and compliance monitoring can improve operational visibility.
The Challenge
High-Volume Vehicle Movement
Thousands of vehicles move through the facility every day, making manual vehicle identification and monitoring difficult to manage consistently.
Vehicle Authorization Checks
Existing cameras captured vehicle movement, but additional intelligence was required to identify whether vehicles were authorized or unauthorized.
PUC Compliance Monitoring
Checking vehicle compliance manually added another layer to the vehicle verification process.
The Solution
Turn Existing Cameras Into Vehicle Intelligence
Mikshi AI added ANPR Video Analytics to the existing camera infrastructure at JSW Steel Dolvi Works.
The system:
- Automatically detected and recognized vehicle licence plates
- Synchronized with the existing JSW Steel vehicle database
- Identified authorized and unauthorized vehicles
- Checked and flagged invalid PUC status
- Displayed detected licence plate numbers through an integrated LED display
Integration with the existing Mantra FR device was also proposed.
This brought video analytics, database synchronization, compliance monitoring, and IoT integration into a single workflow.
Pilot Results
| Metric | Result |
|---|---|
| ANPR Detection Accuracy | 99.5% * |
| Vehicles Monitored | 17,000+ |
| Vehicles Processed | ~2,500/day |
| Vehicle Authorization | Authorized & Unauthorized identification |
| PUC Compliance | Invalid PUC vehicles flagged |
| Real-Time Output | Licence plate displayed on LED |
Client's Opinion
“The pilot worked really well. The ANPR accuracy gave us confidence in the system, especially at this scale.”
Col. Anand Bhargave
JSW Steel Dolvi Works
What Changed
Before Mikshi AI
- Manual vehicle identification and verification
- Additional effort to check vehicle authorization
- Manual PUC compliance verification
- Limited real-time visibility into vehicle movement and compliance
- Licence plate information was not automatically displayed
After Mikshi AI
- Automated licence plate detection through ANPR Video Analytics
- Automated authorized and unauthorized vehicle identification through database synchronization
- Invalid PUC vehicles flagged automatically
- Real-time licence plate display through an integrated LED display
- Automated vehicle intelligence across ~2,500 vehicles per day
- 99.5% ANPR detection accuracy demonstrated during the pilot
*- It is 99.5% on the cases where number plate is visible on the camera
