Classification analysis requires datasets collected using a Classifier Layout, where the A and B axle sensors are placed in parallel and a known distance apart. From this, MCReport derives vehicles.
Firstly, MCReport performs the complex task of examining the raw data and partitioning groups of sensor hits into likely vehicles. This is based on a number of time and distance parameters determined by MCReport.
The next step is to determine the axle configuration of each vehicle. The first A and B hit pair in the group determines the direction of the vehicle, and the speed of the vehicle, based on the sensor spacing. From the speed, the time between the remaining hit pairs determines the spacing between axles.
The final step is to apply a classification scheme, based on the axle spacings in the vehicle. MCReport offers a choice of standard and special-purpose classification schemes, called OEM Schemes. Other classification schemes can be added to MCReport using user-definable External Schemes.
From here, the set of vehicles can be filtered, and formatted into a vast array of reports.
Event Count analysis treats the raw A and B sensor hits as user-selectable events - usually counts. MCReport refers to the definition of an event as the Count Method, which may be one of the following:
- raw counts,
- counts divided by 2,
- counts divided by a custom factor,
- gaps above a certain length (in seconds), or
- following gaps, defined as a starting gap and a following gap.
Datasets collected using a Count Layout should only be analysed using Event Count reports. Attempting to analyse this type of data as vehicles will produce meaningless results.
Classifier Layout datasets can be optionally analysed with Event Count reports. These may be useful for obtaining approximate counts for periods where one sensor has failed, or for gap analysis.