Create Consistent Lighting Conditions
It is crucial to keep lighting both as constant as possible and sufficiently bright to obtain optimal results. All optical factors that may negatively impact MxAnalytics results should be avoided. The results
may be flawed if this is not possible.
The following factors may negatively affect MxAnalytics:
•Constantly changing lighting
•Reflective surfaces (mirrors, windows, etc.)
The MxAnalytics environment settings provide the option to diminish the adverse effects of disruptive lighting such as sunlight glare or shadows in order to improve the analysis result.
Limit Detection Areas (for heat maps only)
Always mark only the areas in the camera live image that you actually want to evaluate in the heat map using MxAnalytics. Omitting non-relevant areas (for example, window surfaces, walls, ceilings, etc.) reduces the processing power required by the camera, which in turn leads to higher frame rates and improved analysis results.
Limit Object Size
A minimum object size for detection should be specified to ensure that the system only tracks and evaluates movement of the relevant objects inside this detection area.
Use time tables for MxAnalytics
If the camera should only be used for MxAnalytics during specific times (for example, Monday to Friday, 9 a.m. to 12 p.m.), it makes sense to configure the camera to use an appropriate time table. This provides an advantage by freeing up the full processing power of the camera in the remaining time for other (security) functions such as event-controlled recording and makes more efficient use of the storage space on the microSD card.
Reduce Image Resolution
The current resolution of the camera should be kept as low as possible (MOBOTIX recommend using VGA) to achieve the highest possible frame rates for MxAnalytics.
Use a high-contrast underlying surface
The more the objects to be recognized differ in color from the underlying surface or background, the more reliable are the MxAnalytics results. An underlying surface with high contrasts (for example, large black and white checkerboard pattern) further improves analysis results.