Surveillance systems have come a long way. From CCTVs that use VCRs to record hours of video on tape to today’s intelligent systems that use AI to automatically identify security risks and use advanced monitoring stations that can respond instantly to threats. However, there are still many challenges facing these intelligent surveillance systems. Many intelligent security systems still lack accuracy, flexibility, and reliability that is required for residential and commercial sites. A significant amount of effort is needed to establish intelligent surveillance more broadly that can respond to growing global security threats. This effort requires collaboration between technology providers, integrators, monitoring operators, and many other industry participants to design and implement more effective intelligent surveillance systems. 

Partnering with industry experts, Camect has created a set of design best practices that aim to advance intelligent surveillance systems for the next generation. These are the 5 guidelines recommended when designing and implementing surveillance systems:

1. Deploy Open Standards – As surveillance technology evolves, open standards enable surveillance system components to work more effectively together and continually improve. Proprietary camera systems that only work with certain platforms or cloud providers continue to constrain progress. New AI models and camera features can become locked out or excluded from surveillance system designs. Open standards such as Onvif for RTSP streams help ensure that cameras can be upgraded and still remain compatible with the system. Implementing AI systems that are optimized on common hardware architectures that support open standards (Intel, Nvidia) can ensure that AI models can be upgraded with new hardware. Ensuring that these open standards are deployed on sites provides the most extensibility for security system components.

> Camect based systems utilize open standards (Onvif, rtsp) that allow integrators to use almost any IP camera. Camect is also optimized on Intel based platforms that are forward compatible with new hardware releases.

2. Ensure APIs Are Available Between Systems – APIs (Application Programming Interfaces) allow apps to communicate with each other and provide flexibility to upgrade independently. APIs allow the applications to continually evolve and provide flexibility to add new applications. For example, a centralized surveillance system may need to expand to directly trigger alarm responses or integrate with home automation systems. A system may also need to interface with a monitoring service in the future. Ensuring the surveillance system applications have APIs will provide that flexibility to interact with other apps and systems in the future.

> Camect based surveillance systems have open APIs available that can be used to integrate with monitoring services, alarm systems, or home automation platforms.

3. Design for Robust Security and Data Privacy – Many surveillance systems are now cloud based and can be susceptible to hacking. This has exposed private data and added new security threats directly to residential and commercial customers. A system should be designed that ensures robust security and minimizes data privacy exposure. Implementing systems that store video onsite and can secure alert data that is sent to the cloud can reduce hacking and privacy risks. Using security standards such as TLS (transport layer security) for data communication within and across applications when a security system is designed can help ensure a robust implementation.

> Camect based surveillance systems store video data privately on-site and only send secure alerts and metadata directly to remote devices.

4. Automate with AI and Maximize Accuracy – Implementing AI at sites that can recognize certain objects and rule out non-threats can automate security operations and save valuable time and resources. However, the myriad of AI solutions available are not all equal. Achieving high accuracy can be very difficult to achieve. High accuracy requires an AI system that is tested and validated at a large scale and can adapt to the environment where the camera is installed. When designing an intelligent surveillance system ensure to select proven AI platforms that have a track record of accuracy and can be customized to the site.

> Camect has established the industry standard for high accuracy and is currently deployed and relied on by professional integrators in over 30 countries. Camect’s AI can be customized for each site and even each Camera by learning the alert preferences.

5. Implement Edge processing to Reduce Cloud Dependencies – Cloud based services can add enhanced functions for a security system, but centralizing with cloud services can also add costs, bottlenecks, and reliability risks. Sending video data to the cloud for processing can consume bandwidth and increase costs significantly. For remote locations that rely on cellular connectivity, transferring video to the cloud can be 9 times more costly than processing locally. Establishing a centralized cloud base security system also can represent a single bottleneck that can fail. If all sites use the same centralized system then the sites may all fail at once and cause a much more significant customer impact. This can be minimized with ‘Edge processing’. Edge processing implements computing devices on site (close to the edge of the network) to handle video processing and other tasks. Combining edge devices that process video locally with a cloud service that shares alert or other metadata with users helps reduce the risks of depending purely on a centralized cloud based video surveillance system.

> Camect’s edge based camera hubs process video locally and are used by many multi-site integrators today to increase reliability and provide accurate alerts.

These best practices are recommended to help design intelligent security systems that allow for continuous improvements. We can expect new hardware capabilities, more advanced camera technologies, and improvements in AI to become available to help improve intelligent security systems. Designing the security systems with open standards, flexible APIs, robust security, high accuracy, and edge processing will provide the best ability to take advantage of these improvements.