Reducing Work Environment Risks With Video Analytics
Overview
Vision AI uses GPU-powered artificial intelligence and deep learning to turn a live video feed into a smart sensor and alerting system that delivers insights to promote safety, reduce theft, and improve customer experiences.
Intelligent video analytics (IVA) allows the system to be trained to identify anomalies that signal safety hazards, security issues, and quality problems.
CBT has both the information technology and operational technology capabilities needed to build the Vision AI video analytics platform from scratch, create a supporting infrastructure, and tailor it to your business needs.
See how Vision AI can be used to monitor your work environment intelligently to detect and prevent risk.
.jpg "Overview")
NVIDIA Tesla GPU
HPE Edgeline Systems to process all camera feeds directly at the edge, in real time, and with full benefits of NVIDIA Tesla-accelerated artificial neural networks. Edgeline systems are compact, energy-efficient, ruggedized platforms with a broad range of network connectivity and data acquisition or control options to accommodate almost all edge video processing use cases.
HPE Edgeline
NVIDIA Tesla GPU analyzes all video streams on the network using deep learning and the NVIDIA Metropolis platform for rich metadata extraction. These architectures are used to accelerate data-intensive video analytics workloads that use deep learning and neural networks to reach actionable insights.
How Vision AI Works
Vision AI leverages machine learning to detect and identify risks. Neural network model training and image tuning are used to teach the system to recognize (inference) visual anomalies and deliver instant actionable insights.
Technologies from NVIDIA and HPE work together to detect, classify, and track objects, as well as recognize faces.
Global situational awareness, combined with deep learning data processing, empowers the system to identify unusual or undesirable scenarios and take action.
Using machine learning and IVA, Vision AI can be taught to identify new anomalies in your work environment.
Using machine learning and IVA, Vision AI can be taught to identify new anomalies in your work environment.
The image sensor system uses video capture to keep a record of what happened before, during, and after the incident. An alert is sent out so the operator can view the video, understand root causes of the incident, and resolve the issue. Alert metrics are generated so your company can track trends to better understand the risk landscape and prevent incidents.
.jpg "How Video as a Sensor Works")
Incident Detection
The system is trained using machine learning to identify incidents that are specific to your company’s risk landscape.
Emergency Response
Alerts can be sent based on real-time information so first responders can get to the site quickly.
Worker Safety/Man Down
Real-time alerts expedite response times after accidents or other workplace incidents. In healthcare, dashboard alerts can be created when required PPE is not being worn by employees or visitors.
Regulatory and Compliance
Vision AI keeps a library of video related to incidents for auditing purposes. Video capture software can preserve footage 30 seconds before and after, as well as during an incident. The system keeps a record of the condition of assets as they arrive and depart.
Quality Assurance
Replace the labor intensive, manual processes of clipboards and pens. Utilize video analytics to automate and expedite quality checks with reduced chance of error.
Inventory Management
Optimize inventory management with automated stocking alerts at the warehouse and in-store to maximize production and delivery efficiency.
Queue Management
Count people waiting in line. Feed wait time data into optimization services like mobile applications for expected wait times, waitlists, or digital signage.
Facility Security
The system can be trained to recognize the markings on company uniforms or recognize faces to identify unauthorized personnel. Corporations with high-value assets can protect their investments by monitoring and reacting to security events in near-real time.
Theft
In retail, the system can detect theft at point of sale. Facial recognition can be used to identify repeat offenders and prevent recurrence.
