Mocolo detects presence of things (and people) in the video frame comparing the current frame with a reference frame. Where the curren frame and the reference frame differ, there Mocolo “finds presence”.
The reference frame pictures an empty scene.
When something enters the scene, current frame ad reference frame don't look the same anymore. The zones where they differ are marked with green artefacts and said to “contain presence”.
Here is the raw result of the comparision between current and reference frame. This picture is stored in the presence buffer.
The comparision is controlled by a threshold parameter. It controls the sensitivity of the presence detection process. Above, a well tuned threshold leads to good presence detection.
Presence detection with threshold too low (left) or too high (right).
After the comparison of reference and current frame, Mocolo builds a list of presence centers. The process is based on a grid: presence within each cell is represented by a green artefact - its position represents the barycentre, its radius represents the “amount of presence” in the cell.
Grid size can be modified depending on users' needs.
Mocolo detects motion of things in the video stream by comparing each frame with the previous one. Where the curren frame and the previous frame differ, there Mocolo “finds motion”.
So, motion detection works almost exactly as presence detection.
Motion is detected in form of motion centers, represented by a red artefact: as for presence, its position represents the barycentre, its radius represents the “amount of motion” - the faster an objects is moving, the bigger its motion centers.
An interaction area is a polygon “drawn on the video frame”. Mocolo checks if any presence (or motion) takes place inside the interaction areas and keeps track of areas statuses, ON or OFF - thus interaction areas behave like virtual triggers.
Here, two interaction areas have been defined to match the yellow sheets. When something enters an area, its status becomes ON.