Implementations of feature tracking methods.


Interface to the tracking module. It returns a callable function for tracking features.


Return a callable function for tracking features.


The Lucas-Kanade (LK) feature tracking module.

This module implements the interface to the local Lucas-Kanade routine available in OpenCV.

track_features(prvs_image, next_image, points)

Interface to the OpenCV Lucas-Kanade feature tracking algorithm (cv.calcOpticalFlowPyrLK).


Thunderstorm Detection and Tracking (DATing) module This module was implemented following the procedures used in the TRT Thunderstorms Radar Tracking algorithm ([HMG+04]) used operationally at MeteoSwiss. Full documentation is published in [FGGB21]. Modifications include advecting the identified thunderstorms with the optical flow obtained from pysteps, as well as additional options in the thresholding.


[HMG+04] [FGGB21]

@author: mfeldman

dating(input_video, timelist[, mintrack, ...])

This function performs the thunderstorm detection and tracking DATing.

tracking(cells_id, cells_id_prev, labels, ...)

This function performs the actual tracking procedure.

advect(cells_id, labels, V1)

This function advects all identified cells with the estimated flow.

match(cells_ad, labels)

This function matches the advected cells of the previous timestep to the newly identified ones.

couple_track(cell_list, max_ID, mintrack)

The coupled cell tracks are re-arranged from the list of cells sorted by time, to a list of tracks sorted by ID.