| Problem | Reason | Fix |
| False alarms | Trees or reflections | Try the Dynamic background model (Environment tab). |
| Rain droplets or insects | Keep the camera lens clean from rain droplets or insects. Rain droplets or insects hitting the camera lens directly will cause false alarms. If a spotlight is installed too close to the camera, it will attract insects. |
| Wind or vibrations causes the camera to move | The camera must be stable. |
| Object stands still in the video for a long time | Track Slow Moving Objects / Object Speed option: Tracks an object moving very slowly across the video. It can also be an object moving directly toward or directly away from the camera, which will appear to be moving slowly. If Track Slow Moving Objects is disabled, objects tend to become part of the background and as such will not be tracked. By enabling this option, these objects will be tracked. Default setting: False
|
| No alarm with fence | The All Fences option has probably been selected | If All Fences has been selected, the object must cross through all fences drawn to cause the alarm. |
| Too many people are counted | Person size calibrated incorrectly | Try decreasing the person size - spatio-temporal cube. |
| Not enough people are counted | Person size calibrated incorrectly | Try increasing the person size - spatio-temporal cube. |
| Person classified as shadow | Person wearing dark clothing on light background | The Shadow Sensitivity slider option controls how aggressively shadows (decreases in lighting) are ignored when tracking moving objects.
Increasing the slider (to the right) will allow the video engine to correctly ignore more of the shadow areas. Default setting: Middle (disabled)
|
| Person classified as an illumination increase | Person wearing light clothing on a dark background | The Illumination Sensitivity slider option controls how aggressively any increase in lighting is ignored. An increase in lighting occurs, for example, when car headlights suddenly appears or when an overcast day becomes a sunny day.
Increasing the slider (to the right) will allow the video engine to correctly ignore any increases in illumination. |
| Background (versus foreground) not determined correctly | If an object has a periodic movement with a period greater than the specified time, then that object will always be considered a foreground object (never merged into the background).
The background is determined based on the last Update Background Seconds of video.
The background is updated all the time, not only when the analytics engine starts. | Adjust the Background learning duration (seconds) option. It determines the approximate length of time it will take a foreground object that becomes stationary to merge into the background. Default setting: 15
|
| Objects with similar colour as background are not detected | Need to reset contrast | The Foreground contrast option sets the minimum contrast needed for an object to be detected as a foreground object.
Default setting: Low
If set to a large value, then objects with similar colour as the background will NOT be detected as valid objects to track. The High contrast threshold option is used to determine how far apart an observation can be from the learnt background model before it is detected as foreground. A low value (for example, 5) will make the background model more sensitive to small changes in the pixel intensities.
Default setting: 25 |