02EarthEmbeddings works with the Google AlphaEarth Foundations Satellite Embedding dataset in QGIS. Every 10 m pixel of every year from 2017 to 2025 carries one 64-dimensional unit vector , , that summarises a year of multi-sensor observation (Brown et al. 2025). Mapping then becomes a matter of comparing vectors: places that look alike have a large cosine .
The plugin reads the public Cloud-Optimised GeoTIFFs without an Earth Engine account and carries a study from the embeddings to results you can defend: example-based similarity search, label-free spherical clustering with model-selection diagnostics, few-label classification with spatial block cross-validation and an area of applicability, design-based accuracy and area estimates with 95 % confidence intervals, multi-year change with calibrated thresholds, polygon embeddings for planning units, and a self-contained HTML report that names every formula and reference. The same chain runs in two ways: the lab dock (page by page, with charts and styled layers) and the Processing provider (one algorithm per step plus a whole-chain algorithm, for models and batch runs). Only qgis.core, GDAL, NumPy and PyQt are needed.
Why it exists: the QGIS tools that use the embeddings each cover part of a study. EMBED-CD does year-over-year change; AlphaEarth Toolbox needs scikit-learn for clustering and classification and reports random k-fold accuracy, which overstates accuracy on spatially autocorrelated pixels; eo_inference_qgis runs other foundation models; GeoAI includes a downloader inside a deep-learning stack; Earth Engine plugins need an account. Outside QGIS, aef-loader and the R package spatial-tools/alphaearth only search and retrieve data. 02EarthEmbeddings covers the chain from data to defended answer in one place.
Dataset. AlphaEarth Foundations Satellite Embedding V1, annual, 2017 to 2025, 10 m, 64 channels A00 to A63, CC BY 4.0. Sources: Google Cloud (official; anonymous HTTPS), the Source Cooperative mirror (same relative paths and identical values), or a local folder laid out as <year>/<zone>/<file> with an aef_index file (GeoPackage, Parquet or CSV). The tile index is read with GDAL's Parquet driver, with pyarrow, or from the local index, and cached per source, year and 1-degree cell.
Reading. Each channel is a signed byte de-quantised and normalised by equation (1):
Tiles are stored bottom-up (positive pixel height) in UTM zones. They are read only through gdal.Warp, which turns them north-up and mosaics zones into one working CRS. Nodata is in the cube and NaN (or a declared value) in float outputs, never 0 or .
Working CRS. The UTM zone of the AOI centre, or the AOI CRS when it is projected in metres and keep CRS is on. All areas are in square metres, hectares or km; pixel sizes are 10, 20, 40 or 80 m (any size of 10 m or coarser in Processing). Vector inputs in any CRS (including EPSG:4326) are reprojected to the cube CRS before sampling.
Estimate before download. The Data page and Estimate download show pixels, tiles, zones, missing years and the size bytes before anything is fetched. An AOI is limited to pixels per year (about km at 10 m); above that a coarser pixel size is suggested.
The AlphaEarth Foundations Satellite Embedding dataset is produced by Google and Google DeepMind. The sentence is written into every layer's metadata, the report and the Help page.
The lab is a dock on the right with a dark icon rail of eight pages and Help. Long steps run in QGIS background tasks with a progress bar and Cancel; the dock never freezes and no network read runs on the GUI thread. Results go to a 02EarthEmbeddings layer group with a subgroup per run. Settings (source, cache folder, AOI mode, pixel size) are kept in the QGIS settings, never in project files.
Area of interest from the canvas extent, a polygon layer or a rectangle drawn on the map; year chips 2017 to 2025; pixel-size chips 10 / 20 / 40 / 80 m; source chips (Google Cloud, Source Cooperative, local folder or raster layers). Check connection queries the index and shows its tier; the CRS badge shows the working CRS and why. Estimate fills the estimate card (pixels, size, tiles, zones, warnings), then Fetch cube downloads. Open existing cube reads a folder with cube.json or 64-band GeoTIFFs; a raster in geographic degrees (an Earth Engine export in EPSG:4326, say), in non-metre units or in Web Mercator is reprojected on the fly into the UTM zone of its centre (nearest neighbour, whole-metre pixels) so areas and block sizes are in metres, and the status line says so; Load demo cube (synthetic) writes the demo of section 5. KPI cards: years, area in km, valid share and the mean vector norm (expected ; a value far from 1 means the rasters are not AlphaEarth embeddings or were rescaled). The PCA colour composite is added automatically.
Positive and negative examples from map clicks (two map tools, markers coloured by role, a list with remove buttons) or from point or polygon layers. Mode chips mean / max / contrast (equations (3) to (5)); threshold chips Otsu / top % / fixed with a live histogram and a draggable threshold line; minimum mapping unit. KPIs: threshold, matched area in ha, regions, best score; a ranked region list zooms the map on click.
range, sample size and seed. Find k draws the silhouette, Calinski-Harabasz and Davies-Bouldin curves with the chosen highlighted; Run clustering maps the clusters. Cluster cards show the colour, share, area, coherence and the nearest cluster; clusters can be renamed in place and the names carry into the report.
Training layer and class field, or quick labels (pick a class, click the map). Samples-per-class bars warn about imbalance. Classifier chips k-NN / prototype / logistic / random forest (greyed out with the install command when scikit-learn is missing); validation chips (spatial blocks, random or both), block size and folds. When every sample falls in one block the run stops and names a remedy (smaller blocks, samples further apart or random validation); when there are fewer blocks than folds, the folds are reduced to the number of blocks. KPIs: OA (blocks), OA (random), the optimism gap , macro-F1 and the share inside the AOA. A confusion-matrix heat map; probability, margin, DI and AOA layers. The report (and the Processing class table) gives the mapped area of every class and the part of it outside the AOA, so a class that absorbs an unknown land cover shows up by its high outside share.
Choose the class raster, the total and the minimum per class, then Create sample. Size n for this precision sets the total from a target standard error of the overall accuracy and an expected user's accuracy for every class, with the mapped shares of the class raster (equation (21)): means about points at 95 %. A labelling helper steps through the points (previous, next, zoom) and writes the chosen class to ref_class, with progress. Estimate areas fills a table of mapped area versus estimated area 95 % CI, UA and PA with intervals, OA CI, and a bar chart with intervals. Every mapped class needs labelled points: a class whose stratum has none gets no estimate (an empty cell, not 0), the estimate covers only the labelled strata, and the status line gives the mapped area left out.
Two or more years, threshold chips (Otsu, MAD, stable reference), an optional stable-mask layer and the minimum mapping unit. KPIs: changed area (ha and %), threshold, the year with most change; a distance histogram; distance, change-year and change-region layers and a before/after PCA pair. Both images of the pair use one PCA basis and one colour stretch, fitted on the two years pooled, so a pixel whose embedding did not change keeps its colour and a colour difference marks a change.
Polygon layer and id field, year, minimum valid share. KPIs: zones and mean coherence. Pick a reference zone to write and colour sim_to_ref; Export similarity matrix (CSV) writes the zone-by-zone cosine matrix.
Writes the HTML report: data provenance with the tile list, attribution and licence, every method used with its equation (typeset with KaTeX, offline) and reference, the KPIs, tables and charts of each run, and the caveats. Export GPKG and rasters copies every result raster to a folder and writes the vectors into one GeoPackage.
A short guide, method references, the data licence and attribution, the cache folder size with Clear cache, and links to this manual and the repository.
The demo cube is synthetic, so every result can be checked against its planted truth: pixels of 10 m in EPSG:32635 (12 km), years 2022 to 2024 (the dock's demo; the Processing demo defaults to 2022 and 2023), five classes plus one novel class, a pixel block (24 ha) that changes class in the last year (2024 in the dock), and an nodata corner in the last year. Class 4 is made confusable with class 3 ().
class, 10 per class), k-NN, block size 500 m. OA (blocks) is 1.0 on the demo; the AOA flags the novel class (which has no training points) as outside, and about 0.5 % of known-class pixels.sim_to_ref = 1.The same example in the Python console with Processing:
import processing
d = processing.run("zero2earthembeddings:demo_cube",
{"YEARS": "2022, 2023", "OUTPUT_FOLDER": "C:/tmp/demo", "TRAINING": "TEMPORARY_OUTPUT"})
r = processing.run("zero2earthembeddings:earth_embeddings_lab",
{"CUBE": "C:/tmp/demo", "YEAR": 2022, "K": 0, "TRAINING": d["TRAINING"],
"CLASS_FIELD": "class", "BLOCK_SIZE": 500, "CHANGE": True,
"OUTPUT_FOLDER": "C:/tmp/demo_lab", "REPORT": "C:/tmp/demo_lab/report.html"})
print(r["K_USED"], r["OA_BLOCKS"], r["CHANGED_HA"]) # 5 1.0 24.0 on the demo
Provider zero2earthembeddings, 13 algorithms in seven groups. Every algorithm opens the cube through the same staged core API as the dock (a cube folder with cube.json, or 64-band year rasters whose year is read from the file name), copies vector inputs and reprojects them to the cube CRS, writes rasters with GDAL (nodata and band names kept) and vector or table outputs with real field types (integers, doubles, text; NULL where a value is undefined). Outputs are not styled by the algorithms. Year 0 means the latest year of the cube. Parameters marked advanced are under Advanced parameters.
Id: zero2earthembeddings:demo_cube
Writes a deterministic synthetic cube with planted truth (marked SYNTHETIC - not AlphaEarth data in layer names, metadata and the report): a Voronoi layout of five land-cover classes plus one novel class that has no training points, a pixel block that changes class between the last two years, and an nodata corner in the last year. Each pixel is its class prototype plus Gaussian noise of standard deviation (default ), normalised and quantised with the inverse of equation (1), so it is read exactly like AlphaEarth data.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
YEARS | Years | Text | 2022, 2023 | |
WIDTH | Width (pixels) | Number | 400 | advanced min 64, max 4000 |
HEIGHT | Height (pixels) | Number | 300 | advanced min 64, max 4000 |
NOISE | Noise (standard deviation) | Number | 0.35 | advanced min 0, max 2 |
SEED | Seed | Number | 7 | advanced min 0 |
OUTPUT_FOLDER | Cube folder | Folder output | - | |
TRAINING | Training points (synthetic) | Vector / table output | - | |
TRUTH | Truth classes (synthetic) | Raster output | - | optional |
Outputs. Cube folder; training points (field class, 10 per known class); truth class raster (Int16, = nodata); the id of the novel class.
Result keys in the dictionary returned by processing.run: NOVEL_CLASS - novel class id (absent from training).
Id: zero2earthembeddings:estimate_download
Looks up the tiles of the source's index that cover the area of interest and returns the grid in the working CRS, the number of tiles, the UTM zones and the download estimate bytes (0.82 is the measured compressed size of one Int8 band value). Nothing is downloaded. Years or tiles missing from the index are listed in the warnings.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
EXTENT | Area of interest | Extent | - | |
YEARS | Years (for example 2024, or 2018-2024) | Text | 2024 | |
RESOLUTION | Pixel size (m, 10 or coarser) | Number | 10 | min 10, max 10000 |
SOURCE | Data source | Choice | Google Cloud (official) | Options: Google Cloud (official); Source Cooperative mirror; Local folder (<year>/<zone>/<file> + index) |
LOCAL_FOLDER | Local folder (local source only) | File / folder | - | optional |
KEEP_CRS | Keep the AOI CRS when it is projected in metres | Boolean | no | advanced |
Outputs. Pixels per year, width, height, tiles, estimated bytes, working EPSG code, warnings.
Result keys in the dictionary returned by processing.run: PIXELS - pixels per year; WIDTH - width (pixels); HEIGHT - height (pixels); TILES - tiles; EST_BYTES - estimated download (bytes); WORKING_EPSG - working CRS (EPSG code); WARNINGS - warnings.
Id: zero2earthembeddings:fetch_embeddings
Reads the planned tiles and writes one north-up 64-band Int8 GeoTIFF per year (aef_<year>_<res>m.tif, bands A00 to A63, nodata , DEFLATE, 512-pixel tiles) and cube.json. AlphaEarth tiles are stored bottom-up (positive pixel height); they are read only through gdal.Warp, which turns them north-up and mosaics several UTM zones into the working CRS. A missing year or tile becomes nodata, never zeros. Pixel sizes above 10 m read Google's averaged overviews. The output folder is the Cube folder of every other algorithm.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
EXTENT | Area of interest | Extent | - | |
YEARS | Years (for example 2024, or 2018-2024) | Text | 2024 | |
RESOLUTION | Pixel size (m, 10 or coarser) | Number | 10 | min 10, max 10000 |
SOURCE | Data source | Choice | Google Cloud (official) | Options: Google Cloud (official); Source Cooperative mirror; Local folder (<year>/<zone>/<file> + index) |
LOCAL_FOLDER | Local folder (local source only) | File / folder | - | optional |
KEEP_CRS | Keep the AOI CRS when it is projected in metres | Boolean | no | advanced |
OUTPUT_FOLDER | Cube folder | Folder output | - | |
PCA | PCA colour composite (latest year) | Raster output | - | optional |
Outputs. Cube folder; years written; width and height; optional PCA composite of the latest year.
Result keys in the dictionary returned by processing.run: YEARS_WRITTEN - years written; WIDTH - width (pixels); HEIGHT - height (pixels).
Id: zero2earthembeddings:pca_rgb
Principal components (equation (2)) of a random sample of valid unit vectors; the first three scores of every pixel are stretched between their 2nd and 98th percentiles and written as red, green and blue (values 1 to 255, 0 = nodata). Similar embeddings get similar colours. The explained-variance shares are stored in the raster metadata item PCA_EXPLAINED.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
CUBE | Cube folder (with cube.json) | File / folder | - | optional |
CUBE_LAYERS | or 64-band year rasters | Raster layers (several) | - | optional |
YEAR | Year (0: the latest year of the cube) | Number | 0 | min 0, max 2100 |
N_SAMPLE | PCA sample size (pixels) | Number | 50000 | advanced min 100 |
SEED | Seed | Number | 42 | advanced min 0 |
OUTPUT | PCA colour composite | Raster output | - |
Outputs. 3-band Byte raster; explained-variance shares of PC1 to PC3.
Id: zero2earthembeddings:similarity_search
Scores every pixel by cosine similarity to example locations, with the mean prototype (equation (3)), the nearest example (equation (4)) or the contrast with negative examples (equation (5)). Points contribute the pixel under them and polygons all their valid pixels. The score is cut by Otsu's threshold (equation (6)), a top percent, or a fixed value; for the contrast score Otsu is applied to the pixels with only, because already prefers the negative prototype. Patches smaller than the minimum mapping unit are removed (8-connected sieve) and the rest are polygonised and ranked by mean score.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
CUBE | Cube folder (with cube.json) | File / folder | - | optional |
CUBE_LAYERS | or 64-band year rasters | Raster layers (several) | - | optional |
YEAR | Year (0: the latest year of the cube) | Number | 0 | min 0, max 2100 |
POSITIVES | Positive examples | Vector features | - | |
NEGATIVES | Negative examples | Vector features | - | optional |
MODE | Similarity | Choice | Mean prototype | Options: Mean prototype; Nearest example (max); Contrast with negative examples |
THRESHOLD | Threshold rule | Choice | Otsu | Options: Otsu; Top percent; Fixed value |
TOP_PERCENT | Top percent (top-percent rule) | Number | 1 | min 0.001, max 100 |
FIXED | Fixed threshold (fixed rule) | Number | - | optional min -2, max 2 |
MMU | Minimum mapping unit (pixels) | Number | 8 | min 0 |
SCORE | Similarity score | Raster output | - | |
MASK | Match mask | Raster output | - | |
REGIONS | Matched regions | Vector / table output | - |
Outputs. Score (Float32, NaN = nodata); match mask (Byte 0/1, 255 = nodata); region polygons with rank, area_ha, mean_score, max_score; threshold, matched area (ha), number of regions.
Result keys in the dictionary returned by processing.run: THRESHOLD_VALUE - threshold; AREA_HA - matched area (ha); N_REGIONS - regions.
Id: zero2earthembeddings:choose_k
Fits spherical k-means (equation (8)) for every in the range on one random sample and returns the cosine silhouette (equation (9), on a 4 000-pixel subsample), the Calinski-Harabasz and Davies-Bouldin indices on chordal distance (equation (10)) and the objective . The chosen has the largest silhouette; ties go to the smaller .
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
CUBE | Cube folder (with cube.json) | File / folder | - | optional |
CUBE_LAYERS | or 64-band year rasters | Raster layers (several) | - | optional |
YEAR | Year (0: the latest year of the cube) | Number | 0 | min 0, max 2100 |
K_MIN | Smallest | Number | 2 | min 2, max 50 |
K_MAX | Largest | Number | 10 | min 3, max 50 |
N_SAMPLE | Sample size (pixels) | Number | 20000 | min 100 |
SEED | Seed | Number | 42 | advanced min 0 |
OUTPUT | diagnostics | Vector / table output | - |
Outputs. Table k, silhouette, ch, db, objective; the chosen .
Result keys in the dictionary returned by processing.run: CHOSEN_K - chosen k.
Id: zero2earthembeddings:spherical_clusters
Spherical k-means (equation (8)) with k-means++ seeding fitted on a random sample of valid pixels; every pixel is then assigned to the most similar centroid, 512 x 512 pixels at a time. With the number of clusters is chosen as in Choose the number of clusters. The profile table gives the pixel count, share and area of each cluster, its coherence (equation (20) over the members) and the nearest other cluster with its cosine.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
CUBE | Cube folder (with cube.json) | File / folder | - | optional |
CUBE_LAYERS | or 64-band year rasters | Raster layers (several) | - | optional |
YEAR | Year (0: the latest year of the cube) | Number | 0 | min 0, max 2100 |
K | Number of clusters (0: choose by silhouette) | Number | 0 | min 0, max 50 |
K_MIN | Smallest (when choosing) | Number | 2 | advanced min 2, max 50 |
K_MAX | Largest (when choosing) | Number | 10 | advanced min 3, max 50 |
N_SAMPLE | Sample size (pixels) | Number | 100000 | min 100 |
SEED | Seed | Number | 42 | advanced min 0 |
MMU | Minimum mapping unit (pixels, 0: none) | Number | 0 | advanced min 0 |
OUTPUT | Clusters | Raster output | - | |
PROFILES | Cluster profiles | Vector / table output | - |
Outputs. Cluster raster (Byte , 255 = nodata); profile table cluster, pixels, share, area_ha, coherence, nearest, nearest_cos; and the silhouette.
Result keys in the dictionary returned by processing.run: K_USED - k; SILHOUETTE - silhouette (cosine).
Id: zero2earthembeddings:classify
Trains a classifier on the embeddings under labelled points (or random valid pixels inside labelled polygons) and maps every pixel: k nearest neighbours with similarity-weighted votes (equation (11), with the smallest class count), nearest prototype with a softmax (equation (12)), multinomial logistic regression (equation (13)) or a random forest (equation (14), only when scikit-learn is installed). Accuracy is estimated twice from the same samples: spatial block cross-validation (square blocks of the block size assigned to folds by a seeded greedy balance of sample counts, every sample of a block in the same fold) and random k-fold, so the optimism gap of equation (15) is visible. Quantity and allocation disagreement (equation (16)) replace kappa. The area of applicability (equation (17)) marks pixels whose embedding is farther from the training data than training samples are from each other across folds.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
CUBE | Cube folder (with cube.json) | File / folder | - | optional |
CUBE_LAYERS | or 64-band year rasters | Raster layers (several) | - | optional |
YEAR | Year (0: the latest year of the cube) | Number | 0 | min 0, max 2100 |
TRAINING | Training points or polygons | Vector features | - | |
CLASS_FIELD | Class field | Field | - | |
CLASSIFIER | Classifier | Choice | k nearest neighbours (cosine) | Options: k nearest neighbours (cosine); Nearest prototype; Multinomial logistic regression; Random forest (needs scikit-learn) |
KNN_K | of k-NN | Number | 5 | advanced min 1, max 100 |
BLOCK_SIZE | Block size of the spatial cross-validation (m) | Number | 1000 | min 1 |
N_FOLDS | Folds | Number | 5 | min 2, max 20 |
SEED | Seed | Number | 42 | advanced min 0 |
AOA | Area of applicability | Boolean | yes | |
PER_POLYGON | Pixels per training polygon | Number | 50 | advanced min 1 |
OUTPUT | Classes | Raster output | - | |
PROBABILITY | Winning-class probability | Raster output | - | optional |
MARGIN | Margin (top-1 minus top-2) | Raster output | - | optional |
DI | Dissimilarity index | Raster output | - | optional |
AOA_MASK | Area of applicability (1 = inside) | Raster output | - | optional |
CLASSES | Class table | Vector / table output | - | |
CONFUSION | Confusion matrix (block CV) | Vector / table output | - |
Outputs. Class raster (Byte class index , 255 = nodata) with the class table value, class, n_train; winning-class probability; margin (top-1 minus top-2 probability); dissimilarity index; AOA mask (1 inside, 0 outside); block-CV confusion matrix (rows = reference); OA (blocks), OA (random), optimism gap, macro-F1, AOA share.
Result keys in the dictionary returned by processing.run: OA_BLOCKS - overall accuracy, block CV; OA_RANDOM - overall accuracy, random CV; OPTIMISM_GAP - optimism gap (random - blocks); MACRO_F1 - Macro-F1, block CV; AOA_SHARE - share of pixels inside the AOA.
Id: zero2earthembeddings:accuracy_sample
Draws a stratified random sample of pixel centres on the classes of a map, the design of Olofsson et al. (2014): every class gets at least the minimum , and the rest of the total is shared out in proportion to the mapped area. With TARGET_SE above 0 the total is computed from equation (21) instead of N_TOTAL. Fill the empty ref_class field with the reference class of each point (the same codes as the map), then run Area estimate.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
INPUT | Class raster | Raster layer | - | |
N_TOTAL | Total sample size | Number | 300 | min 2 |
MIN_PER_CLASS | Minimum per class | Number | 50 | min 1 |
SEED | Seed | Number | 42 | advanced min 0 |
TARGET_SE | Target standard error of OA (0: use the total sample size) | Number | 0 | advanced min 0, max 0.2 |
EXPECTED_UA | Expected user's accuracy (for the target SE) | Number | 0.85 | advanced min 0.5, max 0.99 |
OUTPUT | Accuracy sample | Vector / table output | - |
Outputs. Points sample_id, map_class, empty ref_class in the map CRS; number of points.
Result keys in the dictionary returned by processing.run: N_POINTS - sample points; N_TOTAL_USED - total sample size used.
Id: zero2earthembeddings:area_estimate
Design-based estimates (equation (18)) from the labelled sample: the error matrix in area proportions , the area of each class with its 95 % confidence interval, and overall, user's and producer's accuracy with their intervals. Points with an empty reference, outside the map or on nodata are skipped and counted; a stratum with fewer than two labelled points has no variance and is noted in the log.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
INPUT | Class raster | Raster layer | - | |
REFERENCE | Labelled sample points | Vector features | - | |
REF_FIELD | Reference class field | Field | ref_class | |
OUTPUT | Area estimates | Vector / table output | - |
Outputs. Table per class: map_area_ha, area_ha, area_ci95_ha, users, users_ci95, producers, producers_ci95, n; OA, its 95 % half-width, labelled points used.
Result keys in the dictionary returned by processing.run: OA_CI95 - overall accuracy, 95 percent half-width; N_LABELLED - labelled points used.
Id: zero2earthembeddings:change_detection
The cosine distance between consecutive years (equation (19)). The distance raster holds ; it is thresholded with Otsu (equation (6)), with the median plus MAD (equation (7)), or with the 99th percentile of inside a stable-mask polygon layer. Patches below the minimum mapping unit are removed and the rest are polygonised. The change year is ; it is NaN outside the change mask, so an unchanged pixel has no year. Annual embeddings show that change happened, not when within the year, and an unusual season can raise .
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
CUBE | Cube folder (with cube.json) | File / folder | - | optional |
CUBE_LAYERS | or 64-band year rasters | Raster layers (several) | - | optional |
YEARS | Years (empty: all years of the cube) | Text | - | optional |
THRESHOLD | Threshold rule | Choice | Otsu | Options: Otsu; Median + k MAD; 99th percentile inside a stable-mask layer |
MAD_K | of the MAD rule | Number | 3 | min 0.1, max 20 |
STABLE_MASK | Stable-mask polygons (stable rule) | Vector features | - | optional |
MMU | Minimum mapping unit (pixels) | Number | 10 | min 0 |
DISTANCE | Change distance | Raster output | - | |
CHANGE_YEAR | Change year | Raster output | - | optional |
MASK | Change mask | Raster output | - | optional |
REGIONS | Change regions | Vector / table output | - |
Outputs. Distance (Float32, ); change year (Float32, NaN outside the mask); change mask (Byte 0/1, 255 = nodata); regions area_ha, mean_d, max_d, year_from, year_to; threshold, changed area (ha and percent), the year with most change.
Result keys in the dictionary returned by processing.run: THRESHOLD_VALUE - threshold; CHANGED_HA - changed area (ha); CHANGED_PCT - changed share (percent of valid pixels); YEAR_MOST_CHANGE - year with most change.
Id: zero2earthembeddings:zone_embeddings
The polygon embedding of equation (20): the normalised mean of the valid unit vectors inside each polygon (fields e00 to e63) and the coherence , 1 for a uniform zone. Each polygon is rasterised on its own, so overlapping polygons keep all their pixels. The valid share is the number of valid pixels over the polygon area in pixels; below the minimum the embedding and coherence are NULL. With a reference zone, sim_to_ref . The CSV matrix holds the cosine between every pair of zones.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
CUBE | Cube folder (with cube.json) | File / folder | - | optional |
CUBE_LAYERS | or 64-band year rasters | Raster layers (several) | - | optional |
YEAR | Year (0: the latest year of the cube) | Number | 0 | min 0, max 2100 |
INPUT | Zones (polygons) | Vector features | - | |
ID_FIELD | Zone id field | Field | - | optional |
MIN_VALID_SHARE | Minimum valid share | Number | 0.5 | min 0, max 1 |
REFERENCE | Reference zone (id value; empty: none) | Text | - | optional |
OUTPUT | Zone embeddings | Vector / table output | - | |
MATRIX | Similarity matrix (CSV) | File output | - | optional |
Outputs. Zones with e00 to e63, n_px, valid_share, coherence, sim_to_ref; optional CSV matrix; zones with an embedding and the mean coherence.
Result keys in the dictionary returned by processing.run: ZONES_VALID - zones with an embedding; MEAN_COHERENCE - mean coherence.
Id: zero2earthembeddings:earth_embeddings_lab
The whole chain on one cube through the staged core API, then the HTML report: the PCA composite, clusters ( chosen by silhouette when , range 2 to 10), and, when their inputs are given, similarity search from the positive examples (mean prototype, Otsu), classification with block cross-validation and the area of applicability, change between all years of the cube (two or more) and zone embeddings. Every raster is written to the output folder; the report lists every method used with its equation and reference.
| Parameter | Description | Type | Default | Notes |
|---|---|---|---|---|
CUBE | Cube folder (with cube.json) | File / folder | - | optional |
CUBE_LAYERS | or 64-band year rasters | Raster layers (several) | - | optional |
YEAR | Year (0: the latest year of the cube) | Number | 0 | min 0, max 2100 |
K | Number of clusters (0: choose by silhouette) | Number | 0 | min 0, max 50 |
POSITIVES | Positive examples (optional) | Vector features | - | optional |
TRAINING | Training layer (optional) | Vector features | - | optional |
CLASS_FIELD | Class field | Field | - | optional |
BLOCK_SIZE | Block size of the spatial cross-validation (m) | Number | 1000 | min 1 |
CHANGE | Change between the years of the cube | Boolean | yes | |
ZONES | Zones (optional) | Vector features | - | optional |
ZONE_ID_FIELD | Zone id field | Field | - | optional |
OUTPUT_FOLDER | Output folder | Folder output | - | |
REPORT | HTML report | File output | - |
Outputs. Output folder with every raster; the HTML report; , OA (blocks), changed area, the stages run.
Result keys in the dictionary returned by processing.run: K_USED - k; OA_BLOCKS - overall accuracy, block CV; CHANGED_HA - changed area (ha); STAGES - stages run.
The registry core/analytics/__init__.py holds every method with its LaTeX and number; the HTML report shows the same equations with the same numbers. All vectors are unit vectors after equation (1).
AlphaEarth stores each of the 64 channels as a signed byte . The value is nodata; a pixel with any channel at is invalid. After de-quantisation the vector is normalised, so every method works on unit vectors . The inverse used by the demo and the tests is . Real AlphaEarth vectors have before normalisation; the Data page shows the mean norm as a health check. Never average the raw integers (for example with gdaladdo -r average): the coding is not linear. Reference: Google 2026; Brown et al. 2025. Output: unit vectors.
is a random sample of valid unit vectors (default 50 000), centred on its mean; is the share of variance of component . Reference: Jolliffe 2002. Output: PCA RGB raster.
is the normalised sum of the positive examples' unit vectors. Reference: Brown et al. 2025. Output: score raster.
The similarity to the nearest single example; useful when the examples are varied. Reference: Brown et al. 2025. Output: score raster.
and are the prototypes of the positive and negative examples. With the Otsu rule the threshold is taken on the pixels with . Reference: Brown et al. 2025. Output: score raster.
are the shares and the means of the values below and above on a 256-bin histogram. When several bins share the maximum (an empty gap between two modes) the middle of that plateau is used. Reference: Otsu 1979. Output: threshold.
is the median; 1.4826 makes the MAD a consistent estimate of the standard deviation for normal data; default . Reference: Leys et al. 2013. Output: threshold.
Assignment by the largest cosine, centroids as normalised member sums, k-means++ seeding, several starts (the lowest objective wins). Reference: Dhillon & Modha 2001; Arthur & Vassilvitskii 2007. Output: cluster raster.
is the mean cosine distance of pixel to its own cluster and to the nearest other cluster; computed on a 4 000-pixel subsample. Reference: Rousseeuw 1987. Output: k diagnostics.
On chordal distance; and are the between- and within-cluster sums of squares, the mean distance of members to their centroid . Higher CH and lower DB are better. Reference: Calinski & Harabasz 1974; Davies & Bouldin 1979. Output: k diagnostics.
are the most similar training vectors. Reference: Cover & Hart 1967. Output: class, probability, margin.
is the prototype of class ; the temperature turns cosine similarities into probabilities. Reference: Brown et al. 2025. Output: class, probability, margin.
is the unit vector with a constant 1 appended; ; zero start, full-batch Nesterov gradient descent with step , at most 500 iterations or until the change is below . Reference: Bishop 2006. Output: class, probability, margin.
A bridge to scikit-learn's random forest; the option is greyed out with an install hint when the package is missing. Reference: Breiman 2001. Output: class, probability, margin.
Both cross-validations use the same classifier and samples; a large gap means random folds put neighbouring, near-identical pixels in training and test. Reference: Roberts et al. 2017; Valavi et al. 2019. Output: OA (blocks), OA (random), optimism gap.
, and are the diagonal, row and column proportions of the confusion matrix; . Reference: Pontius & Millones 2011. Output: Q, A.
Unweighted, because all embedding axes share one scale; is the mean pairwise chordal distance of the training vectors, and each training sample's is computed against the samples of other folds; , are type-7 quartiles of the training , as in CAST. Reference: Meyer & Pebesma 2021. Output: DI raster, AOA mask.
is the mapped area share of class , the sample count with map class and reference class , the total mapped area. OA, UA and PA and their intervals follow Olofsson et al. (2014), equations 1 to 7 and 10 to 11. Reference: Olofsson et al. 2014. Output: area +/- CI, OA/UA/PA +/- CI.
for each consecutive pair ; with more than two years the total and the endpoint distance are written too. Reference: Brown et al. 2025. Output: distance, change year, change regions.
is the number of valid pixels inside the polygon (Mardia & Jupp 2000, section 2.3.1). Reference: Mardia & Jupp 2000. Output: e00..e63, coherence.
is the mapped area share of class , its expected user's accuracy (a guess made before labelling) and the standard error wanted for the overall accuracy; is rounded up. This is equation 13 of Olofsson et al. (2014); their worked example, , and , gives . Reference: Olofsson et al. 2014. Output: total sample size n.
| Output | Type | Nodata | Values |
|---|---|---|---|
| Cube year raster | Int8, 64 bands | quantised | |
| PCA composite | Byte, 3 bands | 0 | 1 to 255 |
| Similarity score, probability, margin, DI, change distance | Float32 | NaN | (contrast ), , , |
| Change year | Float32 | NaN (also outside the change mask) | year |
| Masks (match, change, AOA) | Byte | 255 | 0 / 1 |
| Clusters, classes | Byte | 255 | , class index (see the class table) |
| Zone fields | double / integer | NULL | e00..e63, , valid share, sim_to_ref |
Results are deterministic for a fixed seed. Rasters are processed 512 x 512 pixels at a time, never as a whole floating-point cube.
Plugin: GPL-2.0-or-later. Data: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). The AlphaEarth Foundations Satellite Embedding dataset is produced by Google and Google DeepMind.
Cite the plugin as: Eminoğlu, Y. (2026). 02EarthEmbeddings - Earth Embeddings Lab (version 0.1.1) [QGIS plugin]. https://gitlab.com/geophilo1/zero2earthembeddings. https://doi.org/10.5281/zenodo.24263654. The DOI 10.5281/zenodo.24263654 always resolves to the latest version; version 0.1.1 has DOI https://doi.org/10.5281/zenodo.24263655. The repository holds CITATION.cff. Cite the data as Brown et al. (2025) with the attribution sentence above.
Repository: gitlab.com/geophilo1/zero2earthembeddings. Issues: work items. Homepage: geophilo.com/zero2earthembeddings.