Development partner
Fogsphere, supplier of workplace safety monitoring platform based on AI & IoT.
Application
AI PPE monitoring
Challenge solved
Process multiple camera streams and run intelligent video analytics on edge servers in real time
Hardware used
Integrating the Axelera Metis AI Accelerator into our Fogsphere AI platform has significantly bolstered our ability to deliver high-performance safety, security and operational intelligence solutions. It enables us to achieve exceptional price performance across a diverse range of industrial and retail use cases. ~Pasquale Giampa, CEO at Fogsphere
The context:
AI for workplace safety
Lack of preventive safety measures results in millions of injuries, deaths and losses every year.
In light of restrictions in resources - such as budget, manpower – and human abilities, companies in all kinds of industries need to more heavily rely on technology to monitor worker safety, react to workplace risks and ensure operational safety compliance.
Thankfully, AI-driven workplace monitoring platforms enable companies to maintain higher safety standards, while reducing manpower, costs and errors related to manual safety monitoring.
The Fogsphere platform includes smart area controls, automated emergency management, behavior monitoring, vehicle monitoring and AI PPE monitoring. All examples of AI for workplace safety to help ensure safe operations and regulatory compliance.
The challenge:
Simultaneously process multiple streams from the camera
Fogsphere wants to upgrade its existing edge computing system, so it can augment live camera footage with real-time data visualization and risk detection. Thus, their computer vision programs need to be able to simultaneously process multiple streams from the camera in real-time - even under challenging lighting conditions. For this, they need fast, energy-efficient and affordable vision processing units.
To showcase the added value of our hardware, we optimized AI PPE monitoring - as AI PPE monitoring has the potential to prevent many workplace incidents. According to scientific research, not making use of available PPE or improper use of PPE accounts for 34% and 13% of occupational accidents.
This video illustrates a typical use case of real-time video analytics: AI PPE monitoring on a construction site.
The solution:
Include the Metis AIPU
Fogsphere's computer vision software needs to monitor the use of safety gear in the workplace and generate automated responses in real time. Their AI platform for workplace safety identifying unauthorized absence or improper use of protective personal equipment, like helmets or vests.
Fogsphere used the Metis PCIe accelerator card to enhance AI capabilities of their edge server based on Intel Core i5 processor. This setup leverages the 200 TOPS available on the Metis AIPU to run the model in the multiple streams in real-time, even under challenging lighting conditions.
Integrating the Axelera Metis AI Accelerator into our Fogsphere AI platform has significantly bolstered our ability to deliver high-performance safety, security and operational intelligence solutions. It enables us to achieve exceptional price performance across a diverse range of industrial and retail use cases. ~Pasquale Giampa, CEO at Fogsphere
The result:
AI PPE monitoring that excells in performance and price
Fogsphere acknowledges the high-performance and exceptional price of Axelera AI's neural processing unit for accelerating computer vision. The AI PPE monitoring solution of Fogsphere now augments live camera footage with real-time data visualization and risk detection.
The Metis platform allows Fogsphere to enable a multitude of other use-cases, ranging from access control and emergency management to behavioral safety, and emerging AIoT applications.
