Comprehensive video analytics automated pipe accounting and improved industrial safety at the Rashmi Metaliks plant

The Rashmi Metaliks metallurgical plant faced a complex set of problems related to manual accounting of pipe products, low levels of control over compliance with industrial safety standards, and the lack of automated monitoring of production processes.

Project background


The Rashmi Metaliks metallurgical plant faced a complex set of problems related to manual accounting of pipe products, low levels of control over compliance with industrial safety standards, and the lack of automated monitoring of production processes. Manual pipe counting in the warehouse led to inventory errors and financial losses. Monitoring the use of personal protective equipment (PPE), presence in hazardous zones, and compliance with labor discipline was done visually and did not cover all areas on a 24/7 basis. There was no system for early detection of emergency situations (fire).
 

The key objective of the project was to minimize financial losses in the production process — the system was required to identify bottlenecks and automate control at all stages of the production line.
 

A comprehensive automated solution was required for pipe product accounting, industrial safety control, personnel management, and logistics.
 

Implementation progress


Specialists from OSMOVIEW LLC developed and implemented a comprehensive video analytics solution based on the OSMOCV platform. The platform was deployed on the customer's server infrastructure in a local network (2 servers). The integrator provided equipment supply, camera installation, and network infrastructure setup.
 

Video streams from 142 IP cameras (including PTZ cameras for overview control) were fed into the platform, installed in key production areas: pipe warehouse, loading areas, production lines (Main line, SDP3, SDP4, SDP5), entry/exit gates, and hazardous zones. Analytical counting systems were placed between all sections of the production line, allowing deviations to be detected at each stage.
 

The solution included the following in-house neural services developed by OSMOVIEW LLC:

  • Pipe accounting — counting pipes (good/defective) with classification by diameter (350 mm, 400 mm, 450 mm, etc.), detection of stacking violations, measurement of tonnage and pipe “aging” (downtime without shipment);
  • Industrial safety (EHS) — PPE detection (helmets, vests), fire and smoke detection;
  • Personnel management — counting people in zones, detection of absentees, pose recognition (inactivity/falling asleep), mobile phone detection;
  • Logistics — license plate recognition (LPR) at entry/exit, measurement of truck dwell time on the premises, integration with the Challan Management System via API, face recognition (FRS) at checkpoints;
  • Loading equipment management — forklift detection and reverse motion detection with signal light recognition, control of the number of pipe layers during loading.
     

Additionally implemented:

  • Separation of system operating modes (inventory vs. event-based accounting);
  • Generation of alarm events for safety violations and logistical deviations with automatic notification of responsible personnel;
  • Evidence system — automatic saving of video clips with context (before and after the event) for critical incidents; all media files are provided with integrity hash sums (SHA-256) to ensure the immutability of evidence;
  • Role-based access control (RBAC) system and user action auditing to ensure controlled access to data.
     

System calibration, customer personnel training on platform operation, and user acceptance testing (UAT) were conducted. All cases were successfully signed off by Rashmi Metaliks.

Achieved results (confirmed after UAT)


  • Reduced number of industrial safety violations — the number of workers without PPE at workstations has significantly decreased thanks to automatic monitoring and prompt notifications.
  • Reduced use of mobile phones — detecting violations in prohibited zones allowed improving discipline and reducing staff distraction.
  • Improved logistics efficiency — automatic counting of pipes during loading/unloading, monitoring of truck dwell time on the premises, and reconciliation with waybills reduced trip processing time and lowered the risk of errors.
  • Ensured early detection of emergency situations — fire and smoke detection allows for prompt response to threats, minimizing damage.
  • Optimized personnel management — automatic counting of employees in zones and detection of inactivity/falling asleep allow managers to allocate resources effectively.
  • Established a complete evidentiary base — saving clips with SHA-256 hashes ensures the legal validity of data during incident investigations.
  • Identified errors in pipe accounting — the customer's existing counting systems (automatic controllers, i.e., PLCs) periodically output incorrect values, and there were discrepancies even in the existing automation system. Implementing video analytics allowed solving this issue and improving accounting accuracy.