Paradigm shift at the fence
When it is important to provide perimeter protection for sensitive facilities and external building security, thermal imaging cameras are the most popular choice.
They are highly effective for detecting trespassers – but unfortunately they do not provide visual verification of alarms. But when thermal cameras are combined with multifocal sensor technology, a whole range of new capabilities opens up, in terms of analysis, object tracking and identification of individuals on the perimeter. We spoke with Andreas Wolf, Product Manager Intelligent Video Surveillance at Dallmeier.
Mr Wolf, why is multifocal sensor technology more suitable for perimeter protection than conventional HD or megapixel cameras?
Perimeter protection usually involves very long distances. Panomera® is a camera technology developed by Dallmeier that is particularly suitable for monitoring and safeguarding relatively large areas and long distances. With an ingenious array of optical sensors, this patented system is capable of providing surveillance for the entire area in question, with uniform image resolution, good dynamic response and continuous depth of focus from a single location – that would be impossible for conventional single sensor cameras, and the resolution would not be good enough.
So in order to achieve similar resolution quality, an enormous number of single sensor cameras would have to be used. Quite apart from the costs entailed by additional mounting masts, wiring etc., there would still be another problem: The time and labour required to configure each camera installation point individually by hand would then be considerable. Then you must also consider that perimeter surveillance always requires working with 3D models, in order to be able to classify objects meaningfully. The quality of the analysis naturally depends on the camera and analysis having been configured correctly. If a large number of cameras have to be configured manually and this task is not carried out precisely as prescribed, the analysis cannot function reliably. With multifocal sensor technology, the camera is configured automatically, because 3D is already integrated in the Panomera®, so the problem of reliability associated with manual configuration does not exist.
Another dimension that is only possible in conjunction with MFS technology is foreground observation.
What exactly is meant by foreground observation?
If an alarm is set off, the response by security personnel must be rapid and decisive. Particularly when it comes to perimeter protection, it is advisable to continue monitoring the scene so that further movements by the detected object can be observed. With standard perimeter installations, however, the field of view is limited exclusively to the detection zone, that is to say the cameras have a narrow aperture angle and only a small field of view. So it is not possible to see what is happening in the foreground. With multifocal technology, long distances can be monitored, and an object and its location can be detected with certainty. After an alarm is triggered, an object can be displayed automatically in multiple views, so that observation can be optimal. In this way, it is also possible to work out where the object is coming from, and where it is headed, so that it is not immediately lost from view. Because of the high resolution provided by the patented sensor concept in MFS technology, an individual can be detected consistently.
In short, Panomera® at the fence is actually three systems in one: it offers classic intruder detection at the perimeter, analysis and observation of objects well beyond the detection range by virtue of its larger field of view, and lastly detection and identification of individuals through its high resolution.
To this must be added the high level of availability and reliability of the video analysis through the combination of multifocal sensor technology with thermal imaging. Essentially, this new concept in perimeter security heralds a paradigm shift at the fence!
Why does perimeter protection often rely on thermal imaging cameras?
Thermal imaging provides the basis for a very high detection rate with a very low false alarm rate. Because analyses in the visible wavelength range are vulnerable to considerably more potential sources of interference, such as shadows, trees or bushes, and they often reach their performance limits in the face of environmental conditions such as rain, snow, sandstorms, or when something as mundane as the lighting equipment fails and they can no longer detect an object.
Yet thermal technology alone is also not equal to the task when the contrast between objects and their surroundings is not evident based on the thermal conditions. And there is one critical disadvantage associated with the use of thermal imaging cameras: It cannot provide any high resolution colour images of the scenery, that is to say, beyond a generalised alarm verification function, it does not have the capability of detecting intruders, much less identifying them.
So this is why it is used in combination with multisensor cameras?
Coupling Panomera® with thermal technology yields an ideal combination for perimeter surveillance. The purpose of linking these two technologies is so that the weaknesses of the one can be compensated for by the other in each case.
Video analysis is performed primarily with the thermal video, which is less susceptible to interference. If the thermal technology is no longer equal to the task due to the climatic and thermal environmental conditions, the system automatically switches to analysis in the visible wavelength range. Practically total availability of video analysis was not possible until multifocal sensor technology and thermal technology were combined.
At what point exactly do thermal imaging cameras reach the limits of their capabilities?
Thermal sensors measure the thermal radiation from an object and convert it into a grayscale image. The contrast between an object and the background is decisive. If it is inadequate, the thermal video cannot be analysed. The contrast depends on the absolute temperature difference in the image and the relative temperatures of the immediate surroundings and the object. For the purposes of video analysis, thermal sensors lose their effectiveness when the temperature of the immediate surroundings is practically the same as that of the object and the absolute temperature difference is very large, so that few grey values are available per degree of temperature difference. The resulting grayscale image then yields little or no contrast between the object and its surroundings, which means that the object is not visible, and thus also cannot be analysed.
A large absolute temperature difference and a small relative temperature difference between an object and its immediate surroundings occur particularly frequently in industrial environments and the warm summer months from the late afternoon and lasting until the early morning. Reinforced surfaces, streets and building fronts store so much heat that people on or in front of them disappear.
And this is when MFS technology comes into its own?
That’s right. A function called the “Quality of Video” module is permanently monitoring the quality of the video content to check that it can be analysed using video analysis processes. The Quality of Video module captures various measurement values in the video and uses them to calculate sharpness dimensions, contrast ratios and visibility features, among other things, and in the case of thermal sensors the measured temperature ranges as well. When the thermal technology loses its effectiveness for video analysis purposes, the system uses the Panomera® videos for analysis. Switching takes place automatically, but it is reported as an event.
Are the videos from the MFS systems only used for analysis?
No, the video streams from the Panomera® are simultaneously recorded at the highest possible resolution, recording is either continuous or event-controlled. When an incident occurs, the best resolution for visual verification of the event is activated automatically.
What recording and analysis technology is this based on?
The DVS Analysis Server, a complete system installation incorporating both extensive recording functions and very powerful video analysis functions. Licences for preinstalled video analysis applications such as Intruder, PeopleCounting or Object Counting can be obtained as needed for each camera channel individually. Depending on specification, the video analyses operate with a performance of up to 25 fps and an effective resolution of up to 16 MP.
The DVS Analysis Server is based on SEDOR®, a high-performance, self-learning video analysis system that uses the latest image evaluation algorithms and adapts the system parameters continuously to the prevailing ambient conditions to deliver outstanding analysis results. SEDOR® Intruder has the capability to report unauthorised entry in freely definable areas. The application can recognise whether an object is approaching a facility, which direction it is coming from, and how long it remains in an area. The analysis contains special modules and mechanisms that eliminate camera shake, shadows, reflections from the sun and headlights, thereby reducing false alarms to a minimum.
Would existing systems have to be completely replaced to implement this perimeter concept?
All Dallmeier systems work with open interfaces, so the Panomera® solution could also be integrated in an existing thermal camera system. In that case, a phased introduction would also be possible, in which the multifocal technology would be installed first at important nodal points such as entrance roads and gates. We advise customers thoroughly on this aspect, and prepare corresponding migration concepts, because video analysis in all its forms always depends on good planning!