Reconstructs a connected road/path network (graph) from raw GPS tracks using the RidgeSlide density method
GPS Road Builder reconstructs a connected road/path-network graph from raw
GPS tracks of any kind — forestry haul roads, logistics, sports (running,
cycling), off-road, vessels, OSM traces and more. Its core is RidgeSlide — a
Python/NumPy/Numba density method conceived, generalized, developed and
completed by Alexander Kobyakov (building on Guo et al. 2020 and the Slide
method, with his own corrections and additions): build a density surface,
compact trajectories onto road ridges, extract a centerline skeleton, then
weight, clean and connect the edges. A KDE mode is available for very sparse data. Designed
for large inputs (tens of millions of points) with background processing that
keeps the QGIS UI responsive. Heavy optional accelerators (scikit-image,
numba) are installed on demand from a dialog — no silent auto-install.
Interface in Russian and English.
Plugin Tags