Blog

image description

A complete guide to Video Analytics and how it can help reduce your organisation’s overheads

image description

Video surveillance appications have become more sophisticated and accurate in the past few years. As the limitations, and poor accuracy, of the first analytics functions, video motion detection (VMD) and heuristics (manually configured rules) became obvious, i.e. too many false positive alerts, the security industry turned to innovation in machine learning & deep learning (also known as artificial intelligence (AI).

The result is that today's video analytics have far greater accuracy and are increasingly capable of detecting all kinds of objects. Instead of displaying events on a monitor, computer vision (CV) applications can detect objects and classify them, i.e. a person, face or vehicle, recognise patterns of movement, and even alert security teams when an event requires intervention.

In this article, we'll be taking a close look at video analytics and how it can reduce your organisation's security costs and improve outcomes.

What is Video Analytics?

Video analytics applies algorithms, or filters, to camera images & video streams to find patterns or details in the pixels that represent an object (e.g. Person, Face, Vehicle), features (e.g. a nose, a mask, a truck), and/or behaviours (e.g. loitering, unusual movement, fighting). In simple terms, it's the use of computer vision algorithms to detect and classify objects. The most commonly available object detection analytics are person detection, face detection and vehicle detection.

We used to need a person watching a monitor to count how many people appeared in a video. It used to be an exclusively human function to see if a customer was exiting a store without passing through the checkout aisle.

Now, video analytics applications put those functions in the "hands" of algorithms with far greater accuracy. However, there is a strong correlation between analytic accuracy and the computational cost of running it.

Analytics can recognise faces, objects, people & license plates. They can search and analyse video content to return useful footage, saving security teams time and effort. Best of all, they can also help in areas not traditionally related to security including staffing, product placement and store layout.

For the purposes of the security profession, video analytics are still paired with living, breathing operators ............

Speak to a Video Analytics Expert

What Can Video Analytics Do?

Developments in machine learning & deep learning (AI) have unlocked new possibilities and improved video analytic accuracy. We'll go over a handful of existing use cases in this section, and then note a few potential solutions that continue to create excitement in the field and in our team.

Object Detection

It used to be the case that the most a computer could tell you about a video would be things like its runtime, size, and date recorded. Any and all understanding of computer vision was the job of a human.

For an organisation to count the number visitors coming into a building, they'd need to have control room staff view monitors and have them make paper based estimates. Not only is this time-consuming, it's also possible that cognitive bias could lead to poor accuracy.

Today's analytics spare humans most of the mind-numbing work. Object detection and classification algorithms are a category of analytics that when fed a video stream will compare objects in the video against objects that it's been trained to recognise, i.e. a suitcase, laptop or a tennis racket. look for these parameters, and if it finds enough of them, it will decide the video contains a person walking.feature of modern AI that essentially gives video analytics software the power to recognise specific objects in photos and videos.

Developers around the world have been hard at work expanding this capability to all manner of use cases. Algorithms are able to be trained on datasets for deep learning (analytics process images through 3+ layers of filters or interconnected neurons), which determines what the detected object most likely is:on how to detect or recognise people, vehicles, faces, actions and objects. Object datasets can include ietms like guns, backpacks ..... can now study facial expressions and eye movements to diagnose autism among children; that's an indication as to how powerful this technology can be.


Enhanced Spatial Monitoring

A popular benefit of video surveillance systems is the ability to respond to security breaches and other issues in real-time. Organisations rightfully want the ability to keep a watchful eye on their assets and environment. It's one of the reasons why they spend money on CCTV and hire security staff to watch it.

The technology behind this is closely related to object detection. An IP camera pointed at fixed spaces can be instructed to respond to different situations: cars violating traffic or parking rules, people loitering or attempting access to restricted areas, etcetera.

These advances can drastically improve the performance of an organisation's security teams by reducing the risk of human error and increasing response times via instant notifications.

Speak to a Video Analytics Expert

Touchless Access Control

For organisations whose facilities contain restricted areas, access is often a major source of concern. Valuable assets, processes, and restricted areas require as many security measures as possible — which is where facial recognition can assist.

Facial recognition can be used opens a range of use cases for access control. Access points with video monitoring can be set up to include automated alerts when non-uniformed personnel approach, for instance. Likewise, facial and license plate recognition features are at a point where it's possible to grant instant access to personnel who've been preregistered (in a facial recognition system) or whitelisted (their numblate plate has been authorised for entry), reducing the time it takes to clear authorised employees and visitors.

We discuss this in greater length in an article dedicated to video analytics and access control.

Smarter Investigation

No matter how hard an organisation may try to avoid it, there's always a possibility that security incidents will occur. In these cases, much of the value of a video security system is in its ability to aid in investigation.

Video analytics facilitates faster, smarter, and ultimately more accurate investigations. Hours of security footage can be scanned and summarised in minutes thanks to AI, which improves investigation time and limits the possibility of error due to repetition and fatigue.

This is valuable for businesses that face theft on a regular basis, or have a difficult time working with insurance providers to cover losses. Likewise, this presents opportunities for organisations to conduct faster and more precise analysis using their security footage—ultimately resulting in better security protocols down the line.

Better Business Outcomes

The effect that video analytics can have on non-security outcomes is perhaps the most exciting part of this new technology. Security is just one of many use cases that it can deliver.

Organisations can enjoy better KPI tracking (e.g. headcounts, # of interactions with key desks), wider data collection, and opportunities to better understand how their interior ecosystems work. Customer journeys from arrival to PoS can even be tracked and aggregated with the right programming to give insight into smarter marketing methods.

With this technology, the sky's the limit. Security may be the most critical use case and your entry point, but by no means is it where the benefits end. JD Security can help identify the opportunities and develop a solution that can reduce your security risk and overhead costs.

Why Manual Security Video Operation Falls Short

To fully appreciate the benefits of video analytics software, we have to understand the shortcomings of the outgoing generation of solutions.

Man versus machine

One of the reasons why global AI software is expected to be worth $100 billion by 2025 is because computers have a distinct advantage over people when it comes to repetition in certain tasks.

Traditional video surveillance relied on security personnel. This means it was prone to human error, fatigue, and other limitations. On the other hand, systems that incorporate video analytics leave less room for human error.

Of course, it's important to note here that this is by no means a fault of human security operators. Trained professionals bring wisdom and intuition, and we're still a long way from having fully autonomous security systems. It still takes the right people to run and make full use of even the most advanced video analytics tools.

Speak to a Video Analytics Expert

A picture's worthless 'til it's seen

One good way to compare traditional video security against a system with security video analytics is to consider the difference between a reporter and an analyst. In theory, a reporter would simply tell you the facts of an event or phenomenon. An analyst, on the other hand, would dig deeper and explain things like why? or so what?

Even the most advanced camera is simply that: a camera. The most it can offer is a live or recorded view of remote events—perhaps with a better pair of eyes than the average person, but that's where the buck stops.

In contrast, security video analytics offers intelligence on top of mere fact. It turns your video security system from a window to a watcher: one with a superficial level awareness and the agency to notify your staff when key rules are being violated.

Failures of prevention

For larger facilities, it can be almost impossible to rely on video surveillance in order to react to incidents as they occur. Organisations would either have to spend a ton of money on hiring control room staff, outfit their buildings with a variety of sensors, or leave the responsibility of monitoring a whole building to a handful of people.

It's difficult to monitor all interior spaces, and harder still for even a trio of operators to maintain focus on a wide variety of camera feeds. The man-in-a-chair model for video surveillance is far from optimal now that the technology has advanced. A team of veterans could bring the risk factor down significantly, but you'd be right to say that any amount of further risk reduction should be sought after.

Video analytics steps in to provide an autonomous and reliable form of monitoring. Software can be programmed to recognise highly secure areas, differentiate between uniformed and non-uniformed personnel, match faces to a database of authorised employees, and all manner of other functions in between. In terms of prevention, this integrates seamlessly with a security team's messaging and alert system to provide notifications the moment a breach of protocol happens in front of a camera.

Concerns and Limitations

Video analytics technology isn't perfect. There are a handful of concerns and limitations that organisations should be aware of when deciding whether to make the investment. Manufacturers can make big claims, but it takes an experienced team of security professionals and analysts to implement a system that is fit for purpose and considers your particular requirements and challenges.

Privacy

Consumers value their privacy, and so do governments. Care has to be taken when deploying the technology to make sure that no violations of privacy or data collection laws occur. It is critically important to work with experts with a strong legal support system.

Likewise, it'll be as important as ever to communicate good faith and build trust with your stakeholders. This means staff members as well as visitors.

Accuracy Issues

Video analytics never provides 100% accuracy. False positives and false negatives will affect the accuracy of your anayltics applications. Deep learning and machine learning in computer vision is constantly evolving and accuracy has improved significantly in the last 5 years.

Since false positives and failures of detection are possible, it's important to work with a security team that knows how to integrate the analytics that you've chosen within your operating environment and environmental conditions.

Conclusion

There are many reasons why an organisation would do well to consider a security systems provider with video analytics services. The technology offers faster response times, it can improve your business' understanding of your security environment and visitor behaviour, and it can save you the cost of performing the same functions with a larger security team.

When choosing a security partner, it pays to choose a provider with experience in the latest technology with a proven record for generating results with the tools at hand. JD Security offers the best of both worlds: we deploy veteran personnel who value your organisation's outcomes, and equip them with the cutting edge in security technology.

Contact us today to learn how our solutions can make your organisation smarter and more secure, or visit our blog to learn more about the latest trends and developments in video analytics.

Call us on 1300 556 334 or email sales@jdsecurity.com.au to learn more.

Customers in New Zealand call 0800 345 677 or email sales@jdsecurity.co.nz.

Speak to a Video Analytics Expert

Related Articles

Comments are closed.

Return to Top ▲Return to Top ▲