Adoptive Tracking Algorithm for Convection in Meteorological Data (in R)
tracR is an implementation of an efficient and adaptive algorithm for tracking convection in radar, satellite, and simulated data. It uses the Fourier phase shift method to compute the first guess of the motion and the Hungarian method for optimized matching. The algorithm checks for merging and splitting of the objects based on their size and overlapping with the other objects. The algorithm works with bi-level images; hence the tracks are estimated independent of the objects’ physical prop- erties, such as brightness temperature or reflectivity. "TINT is not Titan" is a python version located at https://github.com/openradar/TINT. Both tracR and TINT packages are an independent adap- tion of the prototype (echo_tracking.R) located at https://github.com/RBhupi/Darwin-Rscripts/