Insight by Dell Technologies and Intel

Computer vision and how to expand capabilities in law enforcement and national security

As agencies implement data strategies, hire chief data officers, and devise new applications to use data more effectively, they are starting to deal with one of the fastest-growing sources of data. Video, that is.

The law enforcement domain exemplifies this trend. The growing population of surveillance cameras and related sensing devices practically form an internet-of-things network of their own. But the volumes and nature of the video data present unique challenges.

For one, video data often contains a small proportion of useful or actionable information. For example, hours of surveillance video of, say, a street, military installation or foreign airport might contain hours and hours of effectively the same photo. The appearance of an intruder, the movement of vehicles or people, or any of a thousand occurrences is likely buried somewhere in those hours. So how to you get to the data you really want, short of having scores of people stare at nearly endless video feeds for hours on end?

Other challenges include how to push algorithms for processing data to the edge, to the devices such that they lighten the load of processing the accumulated data at the core or in the cloud. How do you curate test data sets that will train your algorithms to produce accurate and unbiased results? What sorts of people and talents does your agency need to develop the applications and operate them effectively?

To explore these and other questions about video data analysis in law enforcement and national security, Federal News Network and Dell Technologies convened a panel of federal experts.

Watch the video to get up to speed on the latest thinking in use cases and how to make the best use of video data in an increasingly complex law enforcement and national security environment.

The Agency Status on the Internet of Things

IoT [and video] are the types of sensors and information they’re really looking for in order to most into a twenty-first century capability, whether that be law enforcement or any of the other disciplines. The ability to receive sensor data, to increase situational awareness, to assist in investigations is extremely important.

Applications and Use Cases of the Internet of Things

For us, the issue is, how can you take thousands of hours of video camera footage, and quickly go through it in order to find out if there are two people, say, wandering around and then abandoned a package somewhere. That’s the sort of capability we’re really hoping to deliver to law enforcement – not only in the real-time alerting paradigm but also in the past forensics paradigm.

Building Blocks of New Technologies

One of the things we think is critical is … edge compute. That’s where a lot of the information will come in first. You have to have the right kinds of sensors and communications back. Refining [data] as quick as possible…And also to be able to keep the environment healthy but also able to orchestrate and deploy new things in the future. [You need] a scalable architecture from the thin edge to that multi cloud strategy.

Listen to the full show:

Copyright © 2019 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.