In today's rapidly expanding industrial landscapes—particularly in high-tech manufacturing hubs and beyond—a lack of deep, actionable insight is a core challenge for many enterprises. Though thousands of cameras may be deployed across factory floors, the video feeds typically form isolated data silos. As workers move across surveillance zones, tracking is frequently interrupted, creating blind spots. This fragmentation results in a reactive environment, with safety teams reviewing footage after an incident.

Spingence, an AI solutions company based in Taiwan, has enabled proactive, preemptive intervention by integrating NVIDIA Omniverse™ digital twin technology with NVIDIA Metropolis Multi-Target Multi-Camera (MTMC) tracking. Deployed on the NVIDIA-certified OVX server—ASUS ESC8000A-E13P—powered by eight NVIDIA RTX™ PRO 6000 Blackwell Server Edition GPUs, this solution unifies disparate video streams into a cohesive, real-time spatial framework. The solution empowers industrial safety teams to predict and prevent hazardous situations, fostering smarter, more resilient work environments in AI factories worldwide. The data that the system provides also gives engineers insights that help prevent bottlenecks in workflows, potentially improving efficiency.
 

Building Perception and Predictive Capabilities in Four Stages

The integrated platform that Spingence has designed transforms fragmented data into intelligent, actionable insights through four essential stages:

Stage 1: High-Fidelity Digital Twin Modeling

Leveraging NVIDIA Omniverse, a hyper-accurate 3D replica of the factory is constructed, incorporating precise machinery placements, shelving configurations, and spatial parameters. This foundational layer supports advanced AI perception, enabling simulations that mirror real-world conditions without operational risks.

Stage 2: Cross-Dimensional Spatial Calibration and Data Integration

Using NVIDIA DeepStream tools, multiple camera views are precisely aligned with the Omniverse factory floorplan. This creates a unified coordinate system, laying the groundwork for seamless cross-camera tracking and pinpoint localization—eliminating the silos that plague traditional surveillance.

Stage 3: Personnel Identification and Trajectory Tracking

NVIDIA Metropolis MTMC technology assigns persistent digital IDs to individuals across monitored areas. As workers move between zones, the system maintains identity continuity and logs movement paths in real time, providing a comprehensive view of on-site dynamics.

Stage 4: Electronic Geofencing and Instant Alerts

Virtual boundaries are defined for restricted zones, such as machine operation areas or hazardous storage. When unauthorized entry is detected in a restricted zone, immediate notifications and escalations are triggered, facilitating rapid intervention and reducing accident rates by preempting violations.

Robust Hardware for AI-Powered Cameras and Analytics

Processing the intensive computations required for multi-camera AI and real-time analytics requires powerful hardware. The ASUS ESC8000A-E13P, equipped with dual AMD EPYC™ 9005 Series processors, delivers exceptional performance. Offering support for up to eight dual-slot GPUs, and enhanced by NVIDIA® NVLink® Bridge for scalable interconnects and higher bandwidth, the ESC8000A-E13P excels in workloads involving ray tracing, 3D simulation, and AI inference.

Built on the modular NVIDIA® MGX™ architecture, this server offers over 160 customizable configurations, ensuring future-proof compatibility. It enables seamless scaling from pilot testing to full production, boosting return on investment and accelerating time-to-market. This provides a crucial edge that helps enterprises stay competitive amid AI-driven industrial transformations.

Results: Next-Generation Smart Factory Safety and Efficiency

Spingence effectively uses high-performance ASUS hardware to enable multi-tiered technology that refines human-machine collaboration by embedding proactive safety into core operations—delivering robust protection for employees. Utilizing AI to advance from simple visibility to foresight, surveillance blind spots are addressed, resulting in a safer working environment. Data gathered from the system can also help engineers determine more efficient workflows for production lines.

Comprehensive Safety, Risk Management

Intrusion Detection and Anomaly Analysis: Virtual no-go zones around high-risk areas (e.g., electrical panels, automated machinery, etc.) are monitored continuously for unauthorized access, even after business hours. This enhances compliance, minimizes risk, and provides data-driven insights for refining safety protocols—potentially significantly reducing incidents.

Digitally Optimized Operations for Efficiency

1. Zero-Risk Simulation Validation: Using NVIDIA Omniverse's advanced modeling, proposed layout changes or process adjustments can be virtually tested, avoiding disruptions to live production and ensuring zero-risk optimization.

2. Bottleneck Detection and Workflow Enhancement: MTMC-derived trajectory data reveals inefficiencies in personnel flow, enabling data-backed redesigns of facility layouts, equipment positioning, and traffic patterns to maximize throughput and operational efficiency.