By Adam Pease
The Department of Defense (DoD) announced its selection of OStream as a provider to help manage its computer vision-based defense systems. Up until this point, OStream has been a market leader for data management in large physical world AI systems. This blog discusses the DoD’s choice and what it reveals about the way computer vision works in practice.
The Department of Defense (DoD) announced its selection of OStream as a provider to help manage its computer vision-based defense systems.
Why OStream?
OStream provides solutions for object data transfer and management. While it is now easy for organizations that want to deploy systems like computer vision to acquire physical infrastructure like cameras, this equipment is not always optimized for their internal systems or the surrounding software. These problems can become particularly pronounced for organizations with highly rich and layered technology stacks, as is often the case in military contexts.
OStream has had success meeting these needs for IT departments around the world. Aragon has written about the importance of edge computing in emerging computer vision systems. And while edge computing can be the most efficient way to deploy a highly spread-out computer vision system, it can also present challenges in the form of data throughput. OStream unified all camera data into a single data lake, and it is this capability that is helping them meet the needs of large defense organizations.
Computer Vision and Defense
Computer vision is rapidly emerging as an essential infrastructure component of any modern defense or security system. By recognizing and identifying threats, computer vision can expand an organization’s ability to respond dynamically to a changing situation. The DoD has chosen OStream’s Percept solution, which can integrate any camera into its data management ecosystem.
For buyers like the DoD that have a multitude of devices with different standards and software layers to manage, tools that enable the integration of diverse technical systems are essential. The power of a varied set of edge devices cannot be understated, but it should not come at the cost of performance in data management. It turns out that integration providers go hand-in-hand with the best computer vision services.
Bottom Line
While many computer vision products may seem to buyers to be exactly what they need at first glance, it turns out the difficulty of integrating those solutions into their existing technology stack may be more trouble than it’s worth. As a result, providers like OStream are offering the middleware to support critical, device-rich applications of computer vision, such as defense. The DoD’s choice to adopt Percept holds an important lesson about the way enterprises should approach the adoption of computer vision, which often requires additional support systems to smoothly deploy.