Do you want to protect your buildings and grounds from intruders? Depending on your situation, this can be done in all kinds of ways and with different security measures. But always from the outside in and start signaling the intruder. For example, with a camera system with Video Content Analysis (VCA).

Whether you are at high or low risk, the sooner you detect an intruder, the greater the chance that they can be stopped and apprehended before reaching their goal. Ideally, this should be done at the perimeter of your premises. There are various perimeter detection systems available, such as underground detection systems, electric fence systems for walls or fences, and active or passive infrared detection systems. In this article, we will discuss the possibilities of a camera system with video analysis software to monitor your premises and quickly detect intruders.

What is Video Content Analysis?

Video Content Analysis (VCA) software detects changes in the camera’s view, whether in real-time or not. For example, a moving object on the premises. Besides movement, the analysis software can also determine the size and aspect ratio of the object, making it clear what type of object it is (such as a person, animal, or vehicle). The software can even detect the object’s color and determine how fast and in which direction it is moving. The predefined rules you set determine what the software detects and when an alarm is triggered. The software can be embedded in the camera itself, but VCA can also be available in modern Video Management Systems (VMS).

Perimeter detection with VCA

A camera system combined with VCA is highly effective for quickly detecting intruders on your premises. With content analysis software, you can draw a virtual line along the inside of your perimeter, for instance. You can then specify that any detected object (of a certain size, to exclude animals) that crosses this line should be detected, and you want to receive an alarm. This can be in the form of a pop-up in the VMS, an email, or an audio signal. In addition to an alarm, various follow-up actions can be initiated.

Critical success factors

Choosing the right type of camera is essential for the effectiveness of your VCA terrain detection system. This can be either an optical or a thermal camera. An optical camera gives a normal (overview) image, but in the dark or in bad weather conditions it gives a less good image and therefore more unnecessary messages. Thermal cameras (or thermal imaging cameras) perceive people by the heat they radiate. The camera can detect people or vehicles even in pitch darkness, in fog or through dense vegetation. Because this type of camera only provides a thermal image, people or license plates can never be recognized or identified. Unless you combine thermal cameras with optical ones or apply units that include both an optical and a thermal camera. In order to be able to detect successfully, the VCA software must be properly adjusted and tailored to the local conditions. If you set it too tightly, it can easily lead to too many needless alarms. If the opposite is the case, you run the risk of missing suspicious situations. Setting up the system properly takes a relatively long time. We often adjust the settings in the first period after installation to optimize the operation of the system. That is why we connect the systems implemented by Mactwin to the Mactwin Operation Center (MOC) as standard. The setting/adjustment can then be done remotely from the MOC as much as possible, so that we can not only act quickly, but also keep the costs as low as possible. In addition, it is important to project the cameras in such a way that the entire site is well monitored and there are no gaps. The best way depends entirely on the terrain to be monitored, the characteristics of the chosen cameras (e.g. the angle of view) and the possible presence of obstacles (through which the camera cannot see), etc. The cameras should not (or hardly at all) vibrate. A small vibration can cause a big distortion in the image, especially if the camera is looking a little further away. And that, in turn, has a negative impact on detection security. It is also always useful if you can verify reports (afterwards), so that you know what the cause of an unnecessary report is and can ‘learn’ this from the VCA software. And then this. The VCA software used by Mactwin can distinguish between animals and humans. At least, as long as the animal is not too big…. A sheep or very large dog does not differ enough in size from a human, so that the distinction is not always made. Therefore, make sure that such large animals cannot enter the site!

The future

VCA is based on pixel analysis and is relatively static. With the advent of Artificial Intelligence (AI), much more is possible in the field of video analysis. AI is based on ‘artificial neural networks,’ algorithms that attempt to mimic human brain activity and behavior. In practice, this means you can train the software using example videos and situations.

Thanks to AI software, various objects in video footage can be recognized and detected, such as: a person, boat, car, motorcycle, truck, airplane, mobile phone, notebook, television, suspect, weapons, motorcycle helmet, etc. It is even possible to distinguish between people with or without safety helmets or face masks. AI can also detect if people are maintaining the required 1.5 meters distance in a certain zone. Furthermore, AI can detect queues, such as at a reception desk or checkout. Through learning and training, the accuracy and performance of AI algorithms constantly improve.

VCA is not motion detection

Sometimes video content analysis and motion detection are confused. However, they are two different techniques. While motion detection only detects that some pixels in the image have changed, VCA is much more intelligent and can actually recognize what is happening in the image in certain situations.

Combine with other systems for enhanced security

When a camera system with VCA is combined with other perimeter detection measures, an even higher level of security can be achieved. For instance, it is common to surround a site with fencing. These measures will slow down an intruder, while cameras only detect. By adding other detection methods, such as radar detection or an underground detection system, you can more reliably detect unwanted individuals on your premises.