Concept differences between video analytics and motion detector

Question

What is the main difference between the professional video analytics, in particular, developed by Synesis from conventional motion detectors, embedded now in the majority of network cameras and encoders?

Answer

Conventional motion detection are of limited use indoors. Such a detector has no concept of "object", "trajectory", "attributes of the object," "forbidden situation," etc. The detector generates a signal simply "moving" or "no movement" regardless of the nature of this movement. On the other hand, more than 99% of movements in the field of view of the outer chamber is not of interest to the user.

Objectives of the perimeter surveillance in public places, recognizing situations and intelligent search suggests the use of analytics based on a fundamentally more complex hardware and software technologies. Professional video analytics provides at least 1000 times more computing operations than a conventional motion detector.

Video analytics, in contrast to an ordinary motion detector allows you to:

1. Avoid false alarms caused by the environment (lighting change in the motion of clouds, the movement of shadows of trees, glare, rain, snow, insects), and as a result of camera shake.
2. Take into account the three-dimensional model of the scene and scale objects. For example, video analytics will be accompanied by a small figure in the background, and large object in the foreground.With this object in the foreground will not disintegrate into many small objects. This is important for the formation of exactly one event for each object.
3. Accompany the object in the field of view camera. Razryvovy trajectory for object tracking is not highly desirable, since they relate to false positives and increase the cost of ownership. The sides on the other, the exact trajectory can apply the rules (the forbidden zone the forbidden area, etc.)
4. Classify (identify) a person's behavior based on certain rules. Rules can be applied to the speed of the object, the residence time in the area, the emergence of a new facility next to the trajectory (left object), etc.

It is important that the total cost of ownership increases proportionally to the frequency of false positives due to labor costs of operators, transmission and storage of unnecessary data. In contrast to the professional video analytics, conventional motion detector does not allow to significantly reduce cost of ownership while protecting long perimeters, railway or a "safe city".