GEE Time Series Analyzer & Classifier
Spectra, Sentinel-1/-2 and Landsat time series, and supervised crop classification with Google Earth Engine, directly in QGIS.
A dock with four sections. Spectra: surface-reflectance spectra of points and polygons from the Sentinel-2 and Landsat 8/9 image closest to a chosen date (worldwide, cloud-aware), and from local Sentinel-2 L2A scenes (.SAFE / .SAFE.zip, offline). Time series: Sentinel-1 VV/VH backscatter (ascending and descending), Sentinel-2 and Landsat 8/9 optical indices (NDVI, EVI, SAVI, NDWI, NDMI) for polygons drawn on the map, imported from any layer or file, clicked on a field layer or loaded from a GEE table asset; combined, per-polygon and comparison charts, monthly Sentinel-2 composites as map layers, CSV / PNG export. Classification: supervised crop classification (Random Forest, gradient tree boosting, CART) from Sentinel-1/-2 time-series composites, in Earth Engine or offline with scikit-learn; accuracy assessment with HTML report, spatial cross-validation, training checks, field-level results, crop rotation analysis, tree viewer; models are saved and can be applied to other years. Export imagery: composites and time-series feature stacks as GeoTIFF. Also available as Processing algorithms. Requires the "Google Earth Engine" QGIS plugin (provides earthengine-api and the sign-in) and a Google Cloud project registered for Earth Engine; pandas and matplotlib (bundled with QGIS on Windows); scikit-learn only for the offline classification.