printer friendly version

Advancing surveillance

Dallmeier examines the key aspects to consider when considering adding video analysis to your surveillance armoury.

The use of advanced video analysis systems for outdoor surveillance has steadily increased over the last few years. The idea behind it is to detect unauthorised people or vehicles entering protected premises as soon as possible and either notify security personnel and/or trigger automatic actions.

Basically, video analysis systems are designed to detect moving objects and, based on the movement pattern and definable rules, trigger an alarm. The problem with this is the fact that objects of interest rarely behave cooperatively, meaning that they do not want to be detected. Hence, the task of an analysis system is to detect objects under the most adverse operating conditions. Such conditions are often characterised by a low range in contrast and high image noise.

As soon as an analysis system has detected an object, it must keep track of the object both during prolonged standstill as well as at very high or very slow speeds of movement.

Demands on a video analysis system

Which requirements should the customer keep in mind when choosing a video analysis system? It is particularly important to have an overall concept, which should include the following considerations: camera and lighting, recording, interfaces and last but not least, parameterisation and the user interface.

Camera and lighting

A high image quality, which mainly depends on the quality of the camera and the available lighting, is a prerequisite for a minimal rate of false alarms. High-quality cameras with adjusted parameters to meet the requirements of outdoor surveillance are able to provide quality images even under adverse conditions. With such cameras it also does not matter whether analogue or IP technology is used because they provide high-quality images in either case.

In darkness the lighting required for a high image quality is usually provided by infrared lamps. Over the past few years, significant progress has been made in this field and meanwhile high-performance LED lamps are available.

Hardware: embedded vs. PC-based

Video analysis is not easy because various software components are necessary in order to achieve optimal results. The more distinctive the modularity of the system, the more likely customisations can be realised. It is much easier to achieve such modularity using PC-based systems rather than embedded systems as PC-based systems offer a more flexible range of applications. Also, it is easier to scale the available processing power and to maintain the software. This makes it possible to effectively react to customer requirements as they change or evolve during the project. Many years of experience have shown that this is often necessary so as to achieve optimal results.

Interfaces and integration

Generally, video analysis systems used in outdoor surveillance are meant to generate alarms which will usually prompt a human response. The most important interfaces of a system are therefore the alarm and the connected live image of the alarm situation, as well as the video recorded before and after the alarm was triggered.

The video analysis system should provide suitable software interfaces, which can easily be integrated into the management system, for accessing recorded videos as well as the live picture. Such interfaces also allows the targeted search for recorded data.

Parameterisation and user interface

With modern analysis systems the complexity of image processing is hidden beneath easy to use user interfaces, making it possible to configure and parameterise a system even without any profound knowledge. Optimal results can only be realised, though, if no major mistakes are made during the project planning.

Quality of a video analysis system in practice

The quality of a video analysis system ultimately depends on various factors. A high-quality video image, highly sophisticated algorithms for a high reliability of detection and a minimum of false alarms through comprehensive plausibility are crucial to the overall quality of the system.

Owing to their complexity, high-class complete systems only a few manufacturers are able to offer them. Although that complicates the situation for the customers, they should emphasise certain aspects when choosing a system:

* Exclusively demand realistic video material for demonstrations of analysis systems.

* Evaluate the quality of a camera image under realistic conditions.

* Demand that own video material with realistic scenes is used for analysis.

* Evaluate the interaction of all system components as well as their handling.

Share this article:

Further reading: