Tips for Optimizing Counting Accuracy
Last updated
Last updated
This guide describes how to achieve high counting accuracy by optimally choosing the BMA's mounting position and subsequently configuring event triggers.
Before planning and mounting, familiarize yourself with the following points:
The mounting parameters in the BMA's .
The functionality of , , and the .
When planning the BMA's location and alignment, consider the following aspects:
Take into account the specifications in the regarding mounting height, distances, and angles.
Plan the system's mounting so that vehicles are fully visible. Especially with large vehicles (e.g., trucks), ensure they are not only partially visible and do not obscure other vehicles. This can be achieved by:
Elevated positioning without obstacles that could obscure vehicles.
Central placement of the BMA between lanes.
Sufficient distance between the BMA and the roadway; vehicles should be fully visible.
A side perspective (an angle of about 15°) allows for better detection of details like wheels, axles, and windows. This significantly improves classification.
Use the Calibration History to analyze track progressions from the last 24 hours. This is especially helpful during changing light conditions like dusk or night.
When positioning Crossing-Lines, keep the following in mind:
The Crossing-Line should intersect all relevant tracks – especially for larger vehicles like trucks, a slightly offset track is possible.
Avoid positioning at the image edge: Objects are often only partially visible there, which limits detection (visually recognizable by tracks bending sharply).
Place the Crossing-Line as much as possible at the end of the track (but before it bends at the image edge). This gives the system enough time for precise classification.
Objects are only classified once they reach a minimum size (visually represented in dark blue). Do not place Crossing-Lines in these areas, unless the object has already been sufficiently classified beforehand.
Only a specific image area is relevant for measurement. This concentrates all of the system's resources on that Focus Zone.
Track merges occur, meaning the system recognizes tracks from different objects as a single track. This can happen, for instance, at the edges of the image.
Open to get an overview of detected and classified vehicle tracks.
Setting a is recommended if: