Changing Mindsets
Network devices have become smart enough today to discover themselves
- By Adam Lowenstein
- May 14, 2024
We have come a long way from the early days of fuzzy analog CCTV systems. During that time, we have had to migrate from analog to digital signals. When IP-based network cameras arrived, they opened a new world of quality and connectivity but also introduced plenty of challenges. Thankfully, network devices today have become smart enough to discover themselves and even self-configure to some degree. While some IT expertise is certainly required, things are much smoother these days. The biggest change is in how fast security cameras and supporting infrastructure are evolving.
Evolution and Innovation at Breakneck Speeds
Higher resolutions, advanced sensor technology, smarter compression, AI-enabled analytics, and AI image enhancement, all continue to evolve each year. It can be difficult to keep up with all the innovations and even more difficult to know the right time to invest when “the next big thing” is surely right around the corner.
Like our smartphones, it is easy to find yourself waiting for enough new features and functionality to justify an upgrade. For security professionals, there can be budget pressure to install a bare-bones (checkbox) security system that demands compromises. The good news is that even basic camera models get increasingly more powerful every year and are now capable of advanced AI processing at the edge.
Processing this AI within the camera provides significant savings by not requiring powerful backend servers and the additional bandwidth needed to decode extra video streams for analysis.
Riding the AI Wave
AI is in use everywhere, and companies are increasingly aware of the features, benefits and savings attainable when AI is used appropriately. For security use cases, AI-enabled cameras harvest a wealth of information in real-time that can notify operators of events that warrant their attention while simultaneously reducing false alarms.
This enables security teams to be proactive by distinguishing between routine movement and potential security threats, versus merely reacting to events that have occurred in the past. Of course, deep forensic searches for past events also save considerable time and resources as well.
For operations and sales, companies have never been hungrier for data and metrics that validate strategy while revealing inefficiencies and areas for improvement. AI-enabled cameras can provide a wealth of valuable data that displays performance, trends and even compliance.
Keeping Pace with Technology
Keeping up with rapidly evolving technology can be challenging for hardware vendors. The time it takes to research and develop a product, and then get it to market can be significant. For example, to provide the latest AI benefits in a security camera, manufacturers can elect to “roll their own” AI SoC (system on a chip) or use a third-party chip specializing in AI.
Some “pros” of creating an AI SoC in-house include the potential for customization and optimization, along with some cost efficiencies when manufacturing at a large scale. The “cons” of such an approach include the excessive cost of R&D, fabrication and testing coupled with a much longer time to market and lack of flexibility. With the current pace of innovation, there is a significant risk that once an AI chip is finally released to market, it might be already obsolete.
To recoup costs, the manufacturer must continue to sell the technology for a fixed amount of time and then start the process over again. For rapidly changing technology such as AI, using a chip from a third-party specialist like Ambarella makes a lot of sense. Pros include significantly lower initial costs, faster time-to-market, and complete flexibility when upgrading components. Using a chip like the CV52 from Ambarella minimizes risk for manufacturers and customers alike while delivering benefits from AI advances in other industries, such as autonomous vehicles, robotics and more.
Changing Perceptions
The perception of AI-enabled security cameras as IoT devices has evolved significantly as modern security cameras are now part of the broader IoT ecosystem. Manufacturers continually develop new features and capabilities to meet the demands and expectations of businesses, pushing the limits of what these cameras can achieve.
No longer stand-alone devices, the integration of IoT technology has transformed AI-enabled cameras into sophisticated tools that offer real-time monitoring, analytics, and integration with other systems such as access control, alarms, lighting and business intelligence platforms. These integrations further enhance a security team’s ability to provide coordinated and proactive security measures.
When considering security cameras as IoT devices, we naturally assume they are monitored and controlled remotely using smartphones or other devices, allowing operators to view live feeds, receive alerts, and even communicate through two-way intercom features regardless of their location.
However, it is not safe to assume any camera, VMS and access control solution can deliver the integration required. That is why it is important to carefully vet solutions from manufacturers and partners of open platforms to ensure maximum compatibility.
Open vs. Closed
When the industry talks about open platforms, it is making a distinction between manufacturers and providers who supply one-size-fits-all or end-to-end solutions, and those who readily partner and collaborate with best-of-breed manufacturers to give customers a complete choice of deploying bespoke solutions. One-size-fits-all solutions frequently place restrictions on what equipment can be used and, as a result, force compromises.
These end-to-end solutions can seem attractive due to their simplicity and low up-front costs, but the trade-offs mean companies find themselves locked in (Hostage as a Service) and unable to take advantage of the latest innovations that might deliver exactly the capabilities customers require. An example of an open platform might be a readily available software development kit (SDK) or application program interface (API) that third-party developers can use to integrate different applications more easily.
Going further, if a camera manufacturer uses a standardized platform like Docker, a developer of specialized AI tools could create a customized Docker application that runs on any Docker-compatible edge device. This is akin to the apps that run on our smartphones. In this case, the phone OS includes the platform, and the apps are guaranteed to just run.
Enhanced Capabilities Demand More Security
As AI-enabled security cameras become more sophisticated and interconnected, concerns over data privacy and cybersecurity also increase. The capacity of these cameras to harvest and transfer valuable data can make them a preferred target for hackers.
As such, we must ensure that these devices have robust security features to shield them from unauthorized access and potential exploits. As the capabilities of IoT devices expand, so do the legal and ethical considerations surrounding their use. Issues such as consent, data retention and the scope of any surveillance are increasingly under scrutiny, impacting how these devices are deployed and used.
It is time to change our mindset about security cameras, what we can expect from them, and how they fit into the larger ecosystem of IoT devices.
Just as we expect our smartphones to continually evolve through automated updates and ongoing app development, a future-proof platform approach lets us take advantage of new applications and capabilities as they become available. The evolution of security cameras into sophisticated IoT devices reflects broader trends in technology and society.
While they offer significant benefits in terms of security and convenience, they also raise important questions about privacy, data protection, and ethical use. As these devices continue to evolve, it will be crucial to balance these considerations to harness their full potential responsibly.
This article originally appeared in the May / June 2024 issue of Security Today.