3 min

The Edge and Video Analytics

Edge computing and computer vision are helping companies across industries to arrive at valuable insights.

Video analytics (or computer vision) solutions are giving rise to some truly futuristic use cases. 

When people encounter it, especially in a consumer setting, they tend to wonder at the technology. Some of the most promising use cases are currently challenging to scale, but with the right edge computing infrastructure and connectivity, this technology is poised to become widespread in the near future. 

In many industries, video analytics is already capable of having a powerful impact. As IT and business leaders explore the possibilities, they should keep some important edge infrastructure considerations in mind.

Five Use Cases for Video Analytics

1. Retail: Video analytics is helping retailers improve scanning accuracy at the point of sale, prevent theft at self-checkout stations and even enable “autonomous” checkout stores where shoppers are automatically charged for what they take out the door. Video analytics offers a way for retailers to make sure, for example, that customers aren’t switching the price stickers on ground chuck and filet mignon. The solution can also detect and prevent theft by cashiers. 

2. Healthcare: In patient rooms, video analytics can help improve patient safety by detecting falls (or even behaviors that might lead to falls). In operating rooms, computer vision can help ensure that surgical teams follow proper protocols and that they do not inadvertently create a disaster by leaving an item such as gauze or scissors inside a patient’s body. Some healthcare facilities are piloting video analytics through simple, almost mundane use cases, such as monitoring hand hygiene. 

3. Manufacturing: At factories, computer vision can greatly enhance quality control. When a new car rolls off the manufacturing line, for example, analytics from video feeds can alert workers on the factory floor to any defects. Video analytics may also help manufacturers find inefficiencies and optimize their processes. 

4. Agriculture: Some farmers are using video feeds from drones, combined with analytics, to detect rodents or insects in crops. By identifying problems early on, farmers can localize them and prevent them from spreading. While latency is usually not as great a concern in agriculture as it is in other industries, edge computing remains helpful in quickly delivering insights. 

5. Education: In K–12 schools and on college campuses, physical security is becoming an increasingly important consideration. The combination of computer vision, analytics and edge computing can help schools detect weapons and deliver real-time information to first responders, enabling them to identify and neutralize potential threats quickly.

Deployment Considerations

All of these use cases are ready for prime time, and each has been put into practice by organizations in the real world. However, they are mostly emergent applications. No large retailer has switched to autonomous checkout at all of its locations, and the few hospitals using video analytics are largely just testing the technology. 

Robust edge infrastructure is extremely important as organizations seek to deploy and scale video analytics solutions. Fiber connectivity is a must-have, and many locations lack it. (Cellular 5G connectivity may also work for some use cases, but these networks are still maturing.) In addition to computing infrastructure, organizations must find ways to get the power and cooling they need at the network edge, protect equipment from hazards such as dust and set up solutions so they don’t create excess noise.

It will be some time before many organizations are using video analytics at scale. But by keeping their eye on the technology — and building out infrastructure at the edge — they can position themselves to take advantage of its benefits.

Story by Tom Leinberger