All current trackers require a detection feeder. It can be useful to progress a tracker manually without a detection feeder (see Multi-Sensor Fusion: Covariance Intersection Using Tracks as Measurements example).
All current trackers and start their __next__ function with:
time, detections = next(self.detector_iter)
# Some tracking logic
I've moved this line to the new Tracker baseclass and added an update_tracker function which takes the time and detections from the detection feeder as an input. With this being a separate function, you can access it directly and bypass the detection feeder which may be easier in instances.
This also fixes issue #883
This is my third iteration of trying to get this right. See #742 and #779 for previous implementations.
All current trackers require a detection feeder. It can be useful to progress a tracker manually without a detection feeder (see Multi-Sensor Fusion: Covariance Intersection Using Tracks as Measurements example).
All current trackers and start their
__next__
function with:I've moved this line to the new
Tracker
baseclass and added anupdate_tracker
function which takes the time and detections from the detection feeder as an input. With this being a separate function, you can access it directly and bypass the detection feeder which may be easier in instances.This also fixes issue #883
This is my third iteration of trying to get this right. See #742 and #779 for previous implementations.