Motion detection is a process of detecting a change in scenery over a predefined threshold value.
The detection itself can be software-based using a digital recorder with motion detection function or hardware-based using a motion sensor. Each solution has its pros and cons. Motion detection can be indicated with an alarm or can trigger an action (record a video). Software detection allows to determine regions in which motion will be detected and depending on software capabilities, designate areas in which motion will be monitored with high-accuracy. An activation threshold and sensitivity can also be adjusted. After the motion is detected, the recorder starts recording.
Advanced motion detection types are available, including:
line crossing detection – detects objects crossing a virtual line; configured depending on the line crossing direction;
intrusion detection – detects objects entering a pre-defined virtual region;
region exiting detection – detects object exiting a pre-defined virtual region;
unauthorized access detection – detects unauthorized access to the virtual region over a pre-defined time, e.g. for 5 seconds; alarm will be activated if anyone remains in the region for over 5 seconds;
crowd gathering detection – detects the gathering of a crowd in a pre-defined region;
fast movement detection – detects quickly moving objects; a speed at which an alarm is triggered can be set by adjusting sensitivity;
unattended object detection;
object removed detection;
scene change detection;
camera covered detection;
Optionally, the events from before the motion detection can be recorded. If the image is recorded in a loop for a pre-defined time, and if the motion is detected, the device continues recording saves the video from before the event. In this mode, the recorder will operate continuously in normal recording mode but will also provide information on the events immediately before the motion was detected.
Another issue is false alarms. False alarms are a result of motion detection which, from a user’s point of view, is not important including precipitation, an animal passing by or interference due to insufficient lighting. Advanced motion detection algorithms can distinguish between a man and a car, triggering the alarm when an object with a specific shape is detected.
Software detection is often limited by lighting or rather a lack of it. A low noise image is required for efficient motion detection. Without lighting or with insufficient lighting, the image will be dark or noisy and the motion detection will not work as intended. Motion detection requires suitable lighting or the use of a passive infrared sensor (PIR).
Any object at a temperature over the absolute zero emits IR radiation invisible to the human eye. Some devices - IR sensors - can detect this radiation and any changes to it and can work both in complete darkness or in well-lit areas, however, high-accuracy motion detection regions are not supported and the sensors cover an entire region only. The motion detection regions or object shapes cannot be defined. Some PIR sensors can detect object size to prevent triggering by animals, however, the error probability is relatively high.