motionRecognition
syntax
<motionRecognition
duration="50"
areaThreshold="200"
cellOccupancyThreshold="2"
cellOccupationTimeout="5000"
occupancyGridWidth="16"
occupancyGridHeight="9"
>
<motion/>
</motionRecognition>
parameter definitions
duration: type="xs:nonNegativeInteger" default="50"
Number of frames to accumulate in the motion history image (MHI). Increase to account for longer gestures.
areaThreshold: type="xs:nonNegativeInteger" default="200"
Minimum Area (in sensor pixels) threshold for connected components in the motion energy image (MEI), the MHI thresholded to make all nonzero pixels white. Reduce to handle smaller connected components (smaller gesture footprints on MHI). Increase to remove noise and outliers.
cellOccupancyThreshold: type="xs:nonNegativeInteger" default="2"
The number of detected gestures in a cell in the occupancy grid needed to output a motion location in getMotionLocations(). Increase to make less sensitive to spurious detections. Decrease to make more sensitive.
cellOccupationTimeout: type="xs:nonNegativeInteger" default="5000"
Time in milliseconds before resetting a cell in the occupancy grid. Once a cell has been reset, it starts aggregating detections again. Once a cell has reported a detection, it no longer aggregates detections.
occupancyGridWidth: type="xs:nonNegativeInteger" default="16"
Width of occupancy grid. The occupancy grid is overlaid onto the camera image. So a 2x2 occupancy grid in a 320x240 camera image aggregates recognized motions into each quadrant of the image.
occupancyGridHeight: type="xs:nonNegativeInteger" default="9"
Height of occupancy grid. The occupancy grid is overlaid onto the camera image. So a 2x2 occupancy grid in a 320x240 camera image aggregates recognized motions into each quadrant of the image.