PickShift
PickShift: Monte-Carlo estimation of positional and area uncertainty of polygons from GCP planimetric biases.
Polygons digitized from historical planimetric data (old maps, aerial photographs) inherit the positional inaccuracies of their source, yet this uncertainty is rarely quantified. PickShift runs a Monte-Carlo simulation that propagates the planimetric bias measured at Ground Control Points (GCPs) to estimate the positional and area uncertainty of any polygon layer. The bias is interpolated (IDW) over the study extent to produce an error surface, sampled around each polygon vertex through buffered zonal statistics, then randomly resampled over N simulation runs (with optional Douglas-Peucker simplification) to derive, for every polygon, its total uncertainty and 95% confidence interval on area. This plugin is a pure PyQGIS/GDAL/NumPy port of the original PickShift Python script: it only uses libraries bundled with QGIS (qgis.core/qgis.gui, osgeo.gdal/ogr, numpy) and does not require geopandas, fiona, shapely, rasterio, rasterstats or pandas. Source code is GPL-3.0-or-later licensed (see LICENSE). See Jautzy et al. (2024), SoftwareX, 27, 101866 for the underlying method.