How to Reduce False Motion Detection Alerts

Motion detection is useful only when an alert means something. If every cloud, tree branch, reflection, passing headlight, or television image creates an event, users eventually stop paying attention. The goal is not maximum sensitivity; it is reliable detection of the movement you care about.

Security Eye provides two important tools for this job: a motion sensitivity setting and a masking area. They work best when combined with sensible camera placement and testing under the actual daytime and nighttime conditions of the scene.

1. Identify what is causing the unwanted motion

Do not immediately lower sensitivity. First review several false events and look for a pattern. Common causes include:

• sunlight moving across a wall or floor;
• headlights, shadows, or reflections from windows;
• trees, flags, curtains, hanging signs, or plants moving in wind;
• rain, snow, insects, or spider webs close to an outdoor camera;
• TV screens, monitors, digital signs, or changing displays;
• pets moving through an indoor scene;
• automatic exposure or infrared illumination changing when lighting changes; and
• traffic or pedestrians visible in a small part of the frame that is not actually important.

Once you know the source, you can decide whether to change sensitivity, mask that part of the picture, or physically improve the camera view.

2. Tune sensitivity with real movement tests

Open the camera settings and go to the Detection tab. Security Eye's manual recommends adjusting the sensitivity, saving the setting, starting monitoring, and then testing how the detector reacts when a person moves through the scene.

Test at least two kinds of movement:

• normal movement that should trigger an event, such as a person walking through a doorway; and
• small or slow changes that should not trigger an event, such as distant movement outside a window or minor background activity.

Change the sensitivity in small steps. A setting that is too sensitive creates noise; one that is too insensitive can miss real events. There is no universal best value because the correct setting depends on camera position, resolution, lighting, distance, and the amount of normal movement in the frame.

3. Mask areas you do not need to monitor for motion

Masking is often more effective than reducing sensitivity for the entire image. Security Eye can ignore movement inside masked parts of the camera frame while continuing to record the full image. That means you can exclude a problem region from motion detection without cropping it out of the recorded video.

Good candidates for masking include:

• a window with moving trees or traffic outside;
• a road at the top or edge of an otherwise useful property view;
• a television or monitor;
• a ceiling fan;
• an outdoor tree or flag that frequently moves; or
• an area where pets are expected to move but people are not the target of detection.

Do not mask so much of the frame that an intruder can cross the important area without entering an active detection zone. The best mask usually removes a specific nuisance rather than most of the image.

4. Improve the camera view before over-tuning the software

Sometimes the correct fix is physical. A camera aimed directly at a bright window will always have more dramatic light changes than a camera placed at an angle to that window. An outdoor camera pointed through branches will always see motion during wind.

Try these placement improvements:

• aim the camera so important movement crosses the frame instead of occurring very far away;
• avoid placing strong lamps, windows, or reflective surfaces directly in the main detection area;
• keep outdoor lenses and housings clean so insects and webs are less likely to dominate the image;
• use stable mounts to reduce vibration;
• avoid pointing the camera at moving vegetation when another angle can cover the same entrance; and
• if possible, reduce extreme backlighting rather than compensating only with motion settings.

5. Test both day and night

A scene that behaves perfectly at noon can behave very differently after sunset. Streetlights switch on, car headlights sweep across walls, infrared illumination changes the image, and indoor lights may be controlled by timers or occupants.

Test motion detection during the periods when monitoring will actually run. For a business that monitors only after closing, daytime testing alone is not enough. For a home entrance, test both bright daylight and darkness.

6. Separate motion detection from recording strategy

Not every camera needs to use motion as the only reason to record. In some locations, scheduled recording may be more reliable because activity is frequent or the scene changes constantly. In other locations, motion-triggered recording can reduce storage use dramatically.

Security Eye can record when motion is detected and continue recording for a configurable period after the last detected movement. That continuation period helps prevent a single event from being split into many tiny clips when a person pauses briefly.

If a camera produces constant motion events because the environment is inherently busy, consider whether scheduled recording plus selective notifications is a better operational design than trying to eliminate every motion trigger.

7. Use a simple tuning routine

1. Start with a reasonable sensitivity setting.
2. Walk through the important part of the scene and confirm that you are detected.
3. Leave the scene empty and observe what triggers events.
4. Mask specific nuisance areas.
5. Retest normal human movement.
6. Repeat under different lighting conditions.
7. Review the first day or two of events and refine the setting rather than changing many parameters at once.

What not to do

• Do not set sensitivity extremely low just to silence alerts. You may also suppress the movement you actually need to detect.
• Do not mask the primary doorway or path. Mask background nuisances, not the subject of interest.
• Do not judge settings from one short test. Lighting and outdoor conditions change throughout the day.
• Do not assume every false alert is a software problem. Camera vibration, network video artifacts, lighting changes, and scene design can all contribute.

For the original Security Eye instructions, see the Configure Detection section of the User Guide.