How to Leverage Deep Learning and AI-Enabled Data Storage for Heavy Video Analysis

Network video recorders’ new analytics capabilities provide business intelligence

Network video recorders or CCTV systems in the past have been seen as a necessary cost of business to ensure onsite security and for fighting crime. But these systems are now capable of providing new benefits for business intelligence and as an invaluable management tool. Video surveillance technology is transitioning from the old, highly pixelated and grainy low-resolution video that can be pulled up and watched at a later date, to today’s advanced systems: ultra high-resolution video, with the ability to intelligently analyze its video footage in real time.

As demand grows for capturing and managing ever larger amounts of data, and for deeper and more immediate analysis of that data, how can our customers best preserve the integrity of that data? It’s more important than ever that each customer understands how to choose the right hard drive for their application needs. The reliability of a hard drive and the data it manages is directly related to how well the drive matches the workload and the customer’s application. There’s a “right drive” for every application — and choosing the best match is key to long-term reliability.

“By the time 2025 comes, we’ll be gathering 163 zettabytes of data per year,” says Andrew Palmer, Seagate manager of Enterprise and Surveillance Products. “What drives that are things like smart cities. So it’s no longer only the four billion people in the world connected to the Internet. It’s the 26 billion sensors that we’ll be installing, which will help us manage and monitor our daily average life.”

As customer needs evolve, how are advanced video analytics changing?

In the video below, Gary Harmer, Hikvision’s Sales Director for the UK and Ireland, describes the evolution of customers’ needs and the advanced capabilities of video recording systems. Hikvision is the world’s leading provider of innovative video surveillance capture and analysis solutions serving applications from banking, hotel, and retail to industrial, traffic, and mobile transport. With a strong focus on R&D, Hikvision advances core technologies of audio and video encoding, video image processing and related data storage, as well as forward-looking technologies such as cloud computing, big data, and deep learning, smart home tech, industrial automation, and automotive electronics.

 

 

Harmer highlights the industry’s shift from primarily write-only applications to ongoing data monitoring, deep learning and analysis for their customers. He notes how Seagate’s SkyHawk AI enables Hikvision systems to read and write data to the drive at very high speeds to keep up with Deep Learning applications.

These AI applications range from tracking foot traffic to improve retail experiences, understanding what customers are looking for and how they want to shop, to saving lives in healthcare.

5 reasons SkyHawk AI is the better tool for Deep Learning applications

Many of our customers now refer to data itself as the lifeblood of their businesses. And to manage that data, Seagate’s Palmer stresses that it’s crucial to deploy the right drive for the right job. For example, there are important differences between standard desktop drives and those that have been specifically designed and tested for the demanding workloads of today’s multi-camera video surveillance systems offering intelligent analytics.

Hikvision uses Seagate SkyHawk AI hard drives for surveillance analytics applications“At Seagate, whatever your application, we’ve built a drive for it. We build a range of hard drives to suit the fundamentally different jobs that a hard drive is asked to do,” he explains. “Take the comparison between a desktop PC hard drive, built to work 5 days a week, 8 hours a day — that works out to about 2000 hours across a year. On the other hand you have a surveillance hard drive working 24 hours, 7 days a week — that’s almost 8000 hours throughout the year. So there are big differences in the way hard drives operate.”

Palmer describes how Seagate’s technology leadership goes beyond the specialized designs of our hard drives to include involvement with our partners in the development of some of their systems.

“Hikvision has a product called Deep Learning, which is an intelligent recording device,” he says. “It views a scene and learns from that scene what’s normal, and therefore can pick up what’s abnormal behavior and quickly flag it up as an issue. Now that causes a different workload on the hard drive, because no longer are you just recording data, you need to also use that data — it becomes operational, immediately required. We’ve brought out Seagate SkyHawk AI to reflect that change in need.”

Harmer agrees.

Actionable data like footfall analysis, heat-mapping for retail environments, thermal safety measurements

Footfall analysis enabled by Deep Learning and Seagate SkyHawk AI

“It’s absolutely critical to have the right drive deployed in the physical system that does the recording,” Harmer says. “The reason it’s critical is there’s a need to provide a much higher level of integrity of the data, bearing in mind that video footage is often used for evidence purposes. So we cannot have packet data lost, or seconds of video being lost.”

“It’s absolutely critical to have the right drive deployed in the physical system that does the recording.”

“Seagate drives enable us to write a very high speed, and also read back the data at a very high speed,” he explains. “The technology’s moved on from being write-only to the drive, and only read back the data on a periodic basis, to a point where we’re simultaneously reading the data to be able to do the Deep Learning analysis of the content. The move to a drive that can cope with very fast read and write speed times is critical to us.”

With Hikvision’s Deep Learning and SkyHawk AI hard drives, Harmer says, customers can glean entirely new levels of valuable, actionable data from these systems. “We’re able to provide management information regarding footfall analysis, heat-mapping for retail environments, or — for critical safety — temperature measurement on thermal imaging devices that can actually mean the difference between saving a life, and not saving a life.”

2018-03-22T18:49:20+00:00

About the Author:

John Paulsen