printer friendly version

Thermal continues to evolve

It wasn’t too many years ago that thermal cameras suddenly made their debut in the commercial market where they offered a different form of surveillance that made it almost impossible for people to hide. Of course, in these early days, thermals were extremely costly, but some estates bit the bullet to be able to effectively protect their perimeters.

Fast forward to today and thermal cameras are still expensive, but much less expensive than they were when first launched. More estates have installed them and there is more security teams can do with them, much of it automated by analytics.

Today thermal cameras are all the rage because they are said to be able to measure the temperature of people from a distance, and even multiple people at a time. How accurate the various cameras are at this task is not for us to discuss in this article, but we are more interested in the thermals’ applicability to perimeters on estates.

Vanessa Tyne.

Vanessa Tyne, senior key accounts manager and team lead at Axis Communications, states that one of the main reasons to consider thermal cameras on the perimeter is because thermals are less sensitive to problems with light conditions, such as shadows, backlight, darkness and even camouflaged objects. “All these conditions can cause a lot of false alarms, and unnecessary response,” she says. “Thermal cameras deliver images that allow operators to detect and act on suspicious activity 24 hours a day, 7 days a week.”

Quintin Van Den Berg.

“Thermal cameras have been used for several applications, be it perimeter detection, temperature screening of machinery, people and long-range detection solutions, and have an extensive reputation in various fields, from military applications to wildlife conservation,” agrees Quintin Van Den Berg, CCTV product manager at Bosch Security and Safety Systems. “The market has evolved rapidly and there are now multiple form factors and fused (optical and thermal) solutions freely available.”

New developments

Jet Zhu, technical manager, Africa Region for Dahua Technology, adds that more intelligence is being built into thermals through advances in artificial intelligence (AI) which will improve perimeter protection by being able to detect fires, humans and filter out all the false alarms created by animals, light, trees and so forth. Going forward, he says the same thermal systems will be able to double their detection distance with new, optimised software.

Jet Zhu.

One of the latest developments in thermal technology from Axis, according to Tyne, is using thermal cameras and analytics for initial detection, and then through integration with other devices, such as radar and PTZ cameras, allow the estate to verify the alert is a real alarm situation and track the intruders until security can apprehend them.

As a bonus, Tyne notes that thermals offer another benefit in today’s privacy-conscious society, they don’t deliver images that allow reliable personal identification. This is especially useful when considering privacy regulations like POPIA as it ensures an organisation will remain safely in compliance, without missing potential threats.

Van Den Berg adds that we can now also see geolocation data being added to thermal camera feeds in integrations, such as with Bosch and software partners like Genetec. Using this solution, users can follow moving objects on a geographical map with the Restricted Security Area (RSA) surveillance add-on in Genetec Security Center.

“Bosch cameras with built-in video analytics generate information on an object’s location, speed and trajectory using the metadata,” he adds. “Moving targets can automatically be displayed on a map so one can assess and respond to threats accurately in a ‘live time application’ This not only increases the productivity of control room operators, but improves the total cost of ownership by reducing the number of hardware devices that need to be deployed.”

More to come

The evolution of thermals is not over. As AI capabilities advance, we will see more software innovations changing the scope and capabilities of thermal cameras. Smoke detection is one feature that will be advanced, according to Zhu, apart from even more accurate detection and classification of people, animals and so forth.

Tyne agrees, noting that the reduction of false alarms is something that is continually on the radar for improvement. “In many instances, guards have been reduced on site thanks to early thermal detection analytics that work by detecting very small heat signatures,” explains Tyne. “There are some AI analytics available that can do human detection at a few kilometres. This type of analytic is server based at this stage, but there is a big move to do AI on the edge. Axis has a portfolio of on-edge analytics that will allow you to do loitering detection as well as intrusion at the same time and according to the requirements of the site.”

And let’s not forget the possibilities the Internet of Things (IoT) introduces.

“With the Internet of Things (IoT) and increasing intelligence, there are new possibilities for connected security devices beyond their traditional uses,” Van Den Berg says. “Smart security devices now act as sensors, gathering information on activity or objects in an area to provide business and live security insights that bring new value to organisations. Our industry must anticipate how these capabilities will change end-user requirements and preferences.”

The visual question

As noted above, thermal cameras can’t do identification. They can be used to confirm that an object on the perimeter is a human being, but that is as far as it goes. If the security team wants more assistance, they need to look at other technologies, such as visible light cameras.

Van Den Berg notes that being able to expand coverage of long perimeters to identify normal or suspicious activity is critical to the security of these applications. However, he says a fragmented approach with radar, laser, or video analytics results in incomplete or insufficient coverage.

“The solution is seen in the potential of our MIC IP Fusion cameras, the combination of their rugged design with built-in Intelligent video analytics that are specifically designed for the most demanding environments,” he explains. “These solutions fuse the metadata of the camera’s built-in optical (visible light) and thermal imager, providing users with full situational awareness – regardless of whether it’s the optical or thermal video stream that’s being watched.

“Metadata from fusion helps users to focus on ‘invisible’ things that need attention. If an event is detected, but is not necessarily visible in the video stream being watched, an alarm is triggered anyway and users can simply click the alarm overlay box. The user will then see the video data related to that event.”

Van Den berg also expects to see more technologies being pushed into drone solutions and driving down the cost of manufacturing.

Dahua has also met this challenge by adding visible light lenses to some of its thermals to provide the security operators with colour images. The company is also working on thermal-on-drone solutions as this provides a more flexible option in certain environments – and a drone can get to an event faster than a response unit and then follow the intruders.

Tyne also mentions a drive to using radar technology in conjunction with thermal and optical imagery. This allows for heightened security applications for estates that are looking for total solutions. The visual view is for verification, which many end users are requiring for cases that may go to court. Once again, she also sees thermal-on-drone technology gaining some traction, but feels that it is still quite costly for the average perimeter security project.

Latest solutions available

From an Axis Communications perspective, Tyne says the company has a portfolio of dual-head cameras that have both thermal and optical technology built in: the Q87 series. Additionally, it has a full range of thermal cameras in varying resolutions and lens sizes, which are complemented by the Axis on-board Guard Suite analytics.

“You are also able to apply Perimeter Defender to the edge, which delivers a more advanced perimeter detection analytic. From a solution perspective, we are also incorporating audio, allowing estates to place IP horn speakers across the perimeter, linked to the thermal and optical cameras and the control room.”

Bosch will focus on the launch of the INTEOX open camera platform this year. Van Den Berg says that to take full advantage of the opportunities that the IoT brings, the safety and security industry must embrace change. “Bosch is leading by example with the launch of a new camera platform.

“INTEOX recognises the fact that security solutions have far more potential than just creating a safe and secure environment. Security devices are actually sophisticated sensors that provide valuable data that can be analysed and used in a host of new and beneficial ways, many of which we can only imagine.”

For more information contact:

• Axis Communications, +27 11 548 6780, [email protected], www.axis.com

Dahua AI technology ranked #1

Dahua Technology’s AI-based Visual Tracker recently won first place in the comprehensive accuracy ranking of the Generic Object Tracking Benchmark (GOT-10k), surpassing other AI companies and top academic research institutions.

The GOT-10k data set is a general object tracking algorithm evaluation data set with international authority. It is published and maintained by the Intelligent System and Engineering Research Centre of the Institute of Automation, Chinese Academy of Sciences (CASIA). Among the data set, the training set is composed of 10 000 video sequences, including 563 target categories and 87 motion patterns.

This algorithm simulates the human eye’s ability in object motion estimation and tracking. It is responsible for outputting the position and distribution of an object in subsequent frames according to its initial position, while adapting to changes in camera angle, environment, influence of obstructions, surrounding disturbances and other factors. It can be widely used in video monitoring analysis, man-machine interaction, intelligent driving, robotics and other fields.

In order to achieve accurate tracking of moving objects in a complex environment, the RSIA team of Dahua Technology innovatively used a set of representative points instead of bounding box for target state estimation, thereby achieving more fine-grained localisation and modelling of object appearance. As for the framework, two parallel branches were used. Combined with a multi-layer aggregation strategy, it effectively obtained detailed structural information of targets and high discriminative power when handling distractors.

Share this article:

Further reading: