Five factors to Make you Re-Evaluate Video Analytics

Video analytics – or video content analysis (VCA) – should be an integral component of almost every automated surveillance system.

However, VCA has delivered mixed results and attracted criticism when it comes to reliability, practicality and affordability. Over-hyped and over-promised in its formative years, the technology has subsequently offered up an underwhelming choice.

On the one hand, video motion detection (VMD) can add simple and cost-effective monitoring to a camera and CCTV network, but delivers variable detection rates and persistent nuisance alarms. On the other, sophisticated intelligent video analysis (IVA) solutions offer more reliable performance, but are often associated with significantly higher costs and more complex and time-consuming setup.

A new generation of VCA products has recently emerged, broadly termed ‘edge-intelligent IVA’. These applications combine the deployment advantages of edge-embedded VMD with the performance of server-based IVA.

But not all such systems offer affordability and practical set-up. So with that in mind, here are five factors to consider when selecting a video analytics solution…

1. The alarming realities… detection accuracy and nuisance alarms

VMD is a basic video analysis system that uses software to identify movement in a video clip. By defining areas in a scene, VMD can trigger recordings or alerts.

However, a major limitation is an inability to distinguish valid subjects from the effects of weather or variable lighting, resulting in numerous ‘nuisance alarms’. VMD is therefore only suitable for benign environments, such as internal areas. VMD can be deployed onto cameras or in a centralised video management setup.

IVA provides a greater level of functionality, with the ability to define security scenarios and rules, as well as offering a varying degree of nuisance alarm filtering. Scenarios can be basic or span multiple zones with conditional rule-based triggers.

Nuisance alarm filtering can extend to weather, variable lighting and shadowing, car headlights and persistent movement in background vegetation or other scene elements. While this can deliver more reliable detection with less nuisance alarms, IVA also demands more powerful processing, with most systems requiring a dedicated IT infrastructure.

So what does this mean in practice? Many large industrial or critical infrastructure sites adopt server-based IVA for its superior performance.

Such sites are typically able to justify the associated upfront and ongoing costs of IT infrastructure, as well as manage the complexity of what is effectively a significant IT project.

Smaller sites typically opt for simpler edge-embedded VMD, although consistent poor performance often leads to alerts being ignored or deactivated. ‘Edge-intelligent IVA’ presents a viable, reliable option for both large secure sites and smaller sites.

2. The true measure… objective performance certifications

One performance benchmark that has driven adoption of more sophisticated IVA and server-based analytics is the i-LIDS® scheme. The Imagery Library for Intelligent Detection Systems (i-LIDS®) is a UK Government initiative to facilitate development and selection of video analytics to meet Government requirements for security operations.

It uses a video library to benchmark video analysis systems against a number of security scenarios based on accuracy. For example, sterile zone monitoring applications that meet i-LIDS performance criteria may be certified as a primary detection system (i.e. sole security measure) or a secondary system (i.e. supporting measure).

When evaluating the wide range of VCA options on the market, i-LIDS provides an objective and globally acknowledged measure of detection accuracy. While this benchmark is recognised principally by secure sites and critical infrastructure facilities, it also offers guidance to other organisations in their selection of a reliable VCA solution.

In terms of performance, i-LIDS primary certification has typically been the preserve of server-based IVA systems, but a very small number of ‘edge-intelligent IVA’ applications have now attained this definitive performance benchmark.

3. The cost AND value… affordability and return on investment

Organisations that opt for VMD often see it as an affordable way of adding functionality to their camera investments. However, the reliability issues can mean that limited value is realised from these systems, particularly in outdoor applications.

Ironically, while affordability will often persuade a smaller site to adopt VMD, the costs associated with nuisance alarms across multiple sites can often exceed those of a larger site. For example, remote alarm monitoring that relies on VMD can cause multiple unnecessary, costly callouts of security personnel.

While IVA presents a more reliable alternative, the cost of these systems is a major factor in the growing frustration with server-based video analytics. The cost of a single IVA software licence can run at upwards of £400 (€450) per channel, with an additional 10-20% maintenance fee often payable annually.

Combine this with the cost of buying and maintaining server hardware and the total bill for a server-based system can be significant – often exceeding the cost of cameras.

Alternatively, licences for new generation ‘edge-based IVA’ applications can be much cheaper than the server-based equivalent – sometimes costing as little as a quarter of the price. And not all of these licences attract annual maintenance fees or licence renewals.

There is also no need to purchase or install server infrastructure and some systems even incorporate innovative deployment aids that make them almost as simple as VMD.

4. The frustration of installation… practicalities of deployment

One of the attractions of VMD systems is simplicity of installation, although this reflects their inherently unsophisticated functionality, rather than design merits. Then again, most advanced IVA systems require significant effort in calibrating and configuring analytics on each camera.

IVA’s enhanced performance depends on an accurate scene interpretation clearly defined security scenarios.

However, most IVA systems accomplish this by presenting installers with a complex interface, often involving 3D editors and mark-up tools, requiring manual addition of scene and scenario detail – for every camera deployed with VCA. This is often an effort-intensive and specialist task.

Furthermore, the IT infrastructure associated with most IVA solutions involves additional installation effort and cost. Installers also cite additional concerns around server hardware, such as hard disk failures and the high cost of ensuring continuity of power.

Others highlight the fact that continual fine-tuning is required by some analytics packages, making IVA systems increasingly impractical for certain surveillance projects. This combination of complexity, risk and affordability has led many installers and system integrators to question the value and practicality of server-based IVA.

‘Edge-intelligent IVA’ can offer extremely reliable detection without the complexity of a server-based system. In addition, the best of this new generation of VCA products offer innovative approaches to installation with more intuitive, streamlined workflow and the automation of effort-intensive tasks.

This can combine application installation, licence activation and automatic calibration of multiple cameras into a simple setup process that requires just a few clicks. ‘Edge-intelligent IVA’ applications can also exploit native camera outputs to deliver alarms into control rooms and to mobile devices.

5. The inevitability of edge…relentless technology developments point to the future

While the merits of edge-embedded analytics over server-based are obvious, the technology required to deliver ‘edge-intelligent IVA’ has also truly come of age. Continued improvements in processing power on cameras and other edge devices are offering the possibility of ever more enhanced performance for edge-embedded applications.

Most IP cameras and encoders primarily utilise the capacity of on-board processors to perform the core video encoding task. But newer chips offer the potential to ‘piggyback’ more powerful and sophisticated applications onto these devices.

So any organisations considering a VCA investment should not only take into account the practical and cost advantages of ‘edge-intelligent IVA’, but also their longer-term technology roadmap. Any investment in server-based infrastructure today should be factored against the flexibility that an expanding range of ‘edge-intelligent’ cameras and encoders offers both today and in future. Large and small sites alike can benefit from the enhanced reliability, simplicity and affordability that ‘edge-intelligent IVA’ presents. Now is the time to re-evaluate the role that VCA can play in automating your organisation’s security – delivering greater utility and value from your camera investments.

Digital Barriers spoke to us at IFSEC 2014 about why they work with Axis Communications:

Free Download: The Video Surveillance Report 2023

Discover the latest developments in the rapidly-evolving video surveillance sector by downloading the 2023 Video Surveillance Report. Over 500 responses to our survey, which come from integrators to consultants and heads of security, inform our analysis of the latest trends including AI, the state of the video surveillance market, uptake of the cloud, and the wider economic and geopolitical events impacting the sector!

Download for FREE to discover top industry insight around the latest innovations in video surveillance systems.

Download The Video Surveillance Report