“One is not exposed to danger who, even when in safety is always on their guard.”
—Publilius Syrus (Roman author, 1st century B.C.)
With statistics related to crime and security breaches being on the rise, globally, use of video surveillance via analogue or IP-based cameras has become commonplace to counter these threats. Yet, there are times when these devices fail to deliver value, especially when the power or network connection either fails or is disabled by miscreants. In a country like India which lacks power and spectrum consistency, running the surveillance system through uninterrupted power supply (UPS) can also prove to be an expensive option.
But wouldn’t it be great if you could deploy a surveillance system that works even in such odd situations? Silvan Innovation Labs’ smart video surveillance system is one such solution that seemingly addresses many of the bottlenecks that can come in the way of ensuring 24×7 security of an organisation’s premises or public places.
What makes it ‘smart’
The Silvan video surveillance system is an innovative mix of hardware, software and analytics that makes the system intelligent. A few features worth taking note of are embedded storage, distributed intelligence across network, optimisation to save power and image capture even in the dark.
Embedded storage. The system is a combination of cameras that are connected wirelessly through a Wi-Fi connection, thus forming a network. In-built power and storage capabilities make the cameras work and store events even when mains power fails.
Apart from this, one of the cameras in this network even doubles as the ‘master’ camera and has an in-built storage capacity to serve as a network video recorder. It also collates the video images captured by other cameras in the network and streams them in a combined way to the visual medium, thus saving on the cost of buying an additional storage device to store the video footage. For some reason, if the master camera malfunctions and stops to work, the other cameras in the network reconfigure automatically and any of them assumes the role of the master camera storing and streaming the video coverage.
“The on-board memory for video capture and storage helps especially in brown-out conditions (when the network fades or gets disrupted) and forensic reconstruction of events in the case of any security breach,” says Giri Krishna, CEO of Silvan Innovation Labs.
Distributed intelligence across network. The cameras are embedded with algorithms that make them capable of detecting a predefined set of events and generate alerts. This intelligence is distributed across the network of cameras, which means each camera can be programmed to detect a unique set of events and generate alerts. For example, a camera in a public place can be programmed to detect and raise an alert if anyone walks into that place with an object, leaves it behind and doesn’t come back within a predefined stipulated time period to claim it. So when all the cameras in a network work in tandem, they are capable of generating alerts for a range of predefined instances that may lead to security breach.
Optical sensors (or cameras) can also be replaced with other sensors such as IR or thermal using the same analytics for alert generation.
He explains: “With the coming of 3G networks, huge bandwidths will soon be available to telecom operators and high-quality video streaming will become a reality. As a result, the operators will soon be looking at applications that enable the use of this available bandwidth. In this scenario, video surveillance as a service may find traction. Operators these days lease out routers and DSL modems to allow users to connect to the Internet. Along with these devices, they can even install a network of cameras for surveillance (which are available at reasonably low cost nowadays) and store the footage at their data centre. Users can have access to this recording post-event or view it in real time.”
Optimised to save power. The cameras of the Silvan video surveillance system are equipped with an on-board battery that provides up to 30 days of power backup. The patented algorithms embedded in the cameras help them conserve power. For instance, in the event of power failure, when the camera is running on the battery, it may choose to switch off a few functions that are not required. Also, the algorithm may put the camera to sleep in the absence of any movement within the surveillance area and bring it up to function when there is movement.
Image capture even in the dark. The surveillance system can capture video images at night without the need for an expensive day/night camera. The cameras have an integrated passive infrared (PIR) sensor along with LED lighting and a standard CMOS sensor. The PIR sensor detects human movement, which triggers the operation of the camera. The in-built LEDs glow to illuminate the scene and provide a clear and full-colour video output. In the event of a power cut too, the camera continues to capture images. Once the network restores, the stored data is streamed to the master camera and the collected feed is then streamed via the master camera to the visual medium.
What’s most innovative
The company has filed four patents on innovations related to the video surveillance system. The patents cover operational analytics that enable optimal utilisation of power, storage and network bandwidth, and distributed video surveillance. Claims Krishna, “Algorithms that we have built for power optimisation and intelligence distribution across the network is something which has not been reported so far by any other company.”
Krishna recalls and shares a couple of challenges that the team faced during the process of product development. “A team of twelve engineers (hardware and software) worked on this technology for over a year and most of the challenges related to hardware and software co-design. Close collaboration between the hardware and software teams was essential. This requirement was addressed by ensuring that all design decisions were taken jointly and all features were analysed and reviewed jointly to understand the impact of each new addition on hardware and software components.”
Another challenge related to the required expertise and skill-set of the team of engineers. “Since the underlying operating system in use is Linux, the biggest challenge was to put in place a team that was well-trained and knowledgeable in embedded Linux. This continues to be a challenge till date,” says Krishna.
Currently, each camera in the Silvan video surveillance system can do only one type of event detection, based on the algorithm that specifies the event for which the alert is to be generated. But going forward, the Silvan team is working on making the cameras smarter by embedding intelligence that enables each camera to handle complex scenarios and generate alerts for up to three events.
In the face of the security threats that surround us, this is surely going to be a solution to look forward to!.
The author is an assistant editor at EFY