By adopting the NVIDIA Metropolis end-to-cloud video platform, Verizon has built a deep learning application to build a safer, smarter, greener city.
Metropolis is NVIDIA's end-to-cloud video platform for building smarter, faster deep learning applications. Before Verizon, nearly 100 companies were using the Metropolis platform.
Verizon is a leading technology company providing reliable network services across the United States. Its Smart Community Division has been working with cities to connect communities and prepare for future urban planning management, such as connecting NVIDIA Jetson-based smart camera arrays to street lights.
David Tucker, director of product management at Verizon's Smart Community Group, said: "LED streetlights can significantly reduce operating expenses. Cities around the world are expanding their lighting infrastructure into smarter city platforms that enable them to connect current and future applications to help save energy. And create a new convenience service."
This array (called a video node by Verizon) leverages Jetson's deep learning capabilities to analyze multiple video streams to find ways to improve traffic flow, secure pedestrians, and optimize urban parking.
Beta testing using proprietary datasets and models generated through neural network training is nearing the end of the US and East Coast. Verizon will soon release details of its business version.
Predict accidents and prevent them from happening
Launched last year, the NVIDIA Metropolis platform includes the tools, technology and support needed to build deep learning applications ranging from transportation and parking management to law enforcement and urban services.
With NVIDIA Jetson embedded computing platforms and servers and data centers with NVIDIA Tesla GPU accelerators, high-performance deep learning inference can be implemented in the terminal.
Verizon's video nodes use Jetson TX1 to collect and analyze the most remote data on the city's network. The supercomputer on this module accelerates the deep learning of the terminal and enables real-time video analysis. All of these terminal calculations mean more efficient, near real-time data analysis, and lower cost streaming and video storage over LTE and Wi-Fi networks.
Video nodes can capture objects such as vehicles, cyclists, pedestrians, etc., and can identify interactions between objects in a near real-time manner, providing city managers with 24/7 data flow, from red light violations to designated Pedestrian movement outside the crosswalk and then to the parking lot indicator.
Tucker said: "Jetson allows us to use the GPU to create a consistent deep learning view from the cloud to the terminal through the full stack."
Jetson-based nodes are able to identify speeding vehicles, riding motion, and handle other real-time tasks of the terminal. Once the data is returned to the cloud, it can be used for predictive analytics.
Tucker explained: “We tried to understand the impact of intersections B and C on several blocks outside of the neighborhood based on events happening at intersection A.â€
Smart street lights for smarter neighborhoods
In Boston and Sacramento, California, Verizon has deployed video nodes on existing streetlight infrastructure. These smart street lights can “communicate†with autonomous vehicles on the road, helping to reduce congestion and ensure the safety of pedestrians and drivers.
As part of its upcoming solution, Verizon's video nodes support the safety of vehicles, pedestrians and cyclists. To strengthen traffic safety and make communities safer and more harmonious, city leaders around the world have increasingly turned to technology.
More than 35 US cities have signed Vision Zero, a global initiative to reduce pedestrian fatalities. The initiative was first implemented in Sweden in the 1990s and has now expanded to Europe, moving towards the United States, and its core is deep learning.
Led Floor Panels,Led Dance Floor Panel,48W Led Light Panel,Floor White Uplight Panel Led
Kindwin Technology (H.K.) Limited , https://www.szktlled.com