Motion detector setup recommendations🔗
The software motion detector analyzes the video stream received from the camera and, in case of detecting motion, generates the corresponding system event, based on which, for example, a recording to archive is started.
The motion detector consumes certain system resources of the server: loads up the processor, uses the RAM memory. At that, it can be estimated that the consumption of system resources is proportional to the resolution and frame rate. For example, for a video stream with a resolution of 1920x1080 with the use of motion detector, the processor load is approximately 6 times higher than for the 640x480 video stream. Likewise, when the analysis is performed with the frequency of 25 fps, the consumption of system resources is 4 times higher than when the analysis is performed with the frequency of 6 fps.
At the same time, triggering the recording by the motion detector allows to reduce the archive volume because the recording is performed only when there is motion in the frame.
To avoid the situation when the software motion detector is triggered by such insignificant movements as those of leaves, birds etc. it is possible to set it up for detecting the movement of the objects bigger that the preset size.
Note
The software motion detector must be enabled on the following analysis modules:
Sabotage detector;
Loud sound detection;
Auto zoom;
Tracking;
Search for objects;
People counting;
Crowd monitoring;
Face detection;
Traffic density heat map;
Personnel activity monitoring;
Uniform Detector;
Shelf Fullness Check;
License Plate Recognition (Complete);
License Plate Recognition (Light).
Many cameras have their own, inbuilt motion detectors (in Eocortex terms — Camera motion detector). When the camera motion detector is used, the system resources of the server are practically not consumed.
If it is planned for the camera to perform continuous recording to archive, and it is not planned to use video analytics modules, the use of the software motion detector is not recommended. Instead, it is sufficient to select the mode of continuous recording to archive. Such situation is possible when, for example, there is motion that happens 24/7: during the continuous manufacturing process, at a large railway station etc.
If the camera is equipped with the inbuilt motion detector, and it is not planned to use video analysis modules, it is recommended to enable the camera motion detector. It allows to lower the server load and reduce disk space usage at the same time.
That said, in some situations it is preferable to use software motion detector, for example, when there are many cameras of the same type and when setting up a motion detector on each camera would be too time-consuming. In such a case it is easier and faster to enable the software motion detector for a group of cameras, or set up one camera and copy its settings many times, changing only the network address.
In Eocortex, it is also possible to set up the combined mode of recording to archive (in Eocortex terms, the Scheduled mode). For example, there is constant movement at the entrance of a shopping mall from 10:00 to 22:00, but at other times there should be no movement. In such a case, it is possible to set up the recording as follows: from 10:00 to 22:00 — continuous recording (this will reduce server load); from 22:00 to 10:00 — motion detection-triggered recording (this will save dick space).
By default, the frequency of checking video stream for the presence of motion in the frame, which is called detection frequency, does not match the frame rate of the video stream: most of the time, the detection frequency is lower than the video stream frame rate. It is done to reduce the consumption of resources by Eocortex server. This mode is called detection with optimal frequency.
At the same time, it is possible to set constant detection frequency in the software motion detector settings in the Eocortex Configurator application. It is advisable to use this mode when the optimal detection frequency is a priori lower or higher than the one required and sufficient for analysis, or when it is necessary to set the fixed frequency for H.264, H.265 or MPEG-4 formats. The analysis frequencies recommended for a particular video analytics module are stated in the documentation for each module.
Note
When multiple streams from a camera are used, the motion is by default detected on the first additional stream.
At the same time, in the software motion detector settings of Eocortex Configurator application it is possible to select which stream will be used for analysis on the camera.
If no other video analytics modules are used except for the software motion detector, it is recommended to use lower resolution stream for detection because it allows to decrease server load significantly.