| { |
| "@context": { |
| "@language": "en", |
| "@vocab": "https://schema.org/", |
| "citeAs": "cr:citeAs", |
| "column": "cr:column", |
| "conformsTo": "dct:conformsTo", |
| "cr": "http://mlcommons.org/croissant/", |
| "rai": "http://mlcommons.org/croissant/RAI/", |
| "data": { |
| "@id": "cr:data", |
| "@type": "@json" |
| }, |
| "dataType": { |
| "@id": "cr:dataType", |
| "@type": "@vocab" |
| }, |
| "dct": "http://purl.org/dc/terms/", |
| "examples": { |
| "@id": "cr:examples", |
| "@type": "@json" |
| }, |
| "extract": "cr:extract", |
| "field": "cr:field", |
| "fileProperty": "cr:fileProperty", |
| "fileObject": "cr:fileObject", |
| "fileSet": "cr:fileSet", |
| "format": "cr:format", |
| "includes": "cr:includes", |
| "isArray": "cr:isArray", |
| "arrayShape": "cr:arrayShape", |
| "isLiveDataset": "cr:isLiveDataset", |
| "jsonPath": "cr:jsonPath", |
| "key": "cr:key", |
| "md5": "cr:md5", |
| "parentField": "cr:parentField", |
| "path": "cr:path", |
| "recordSet": "cr:recordSet", |
| "references": "cr:references", |
| "regex": "cr:regex", |
| "repeated": "cr:repeated", |
| "replace": "cr:replace", |
| "sc": "https://schema.org/", |
| "separator": "cr:separator", |
| "source": "cr:source", |
| "subField": "cr:subField", |
| "transform": "cr:transform", |
| "wd": "https://www.wikidata.org/wiki/", |
| "prov": "http://www.w3.org/ns/prov#" |
| }, |
| "@type": "sc:Dataset", |
| "conformsTo": "http://mlcommons.org/croissant/1.0", |
| "name": "TreeUQ – Bavaria EO Benchmark", |
| "description": "TreeUQ is a large-scale Earth Observation benchmark for tree species classification and tree-structure estimation with uncertainty quantification, covering the federal state of Bavaria, Germany. It contains 45,419 patches (train 31,092 / validation 8,150 / test 6,177). Each 128×128-pixel patch at 10 m GSD contains: four-season Sentinel-2 L2A composites (10 bands), four-season Sentinel-1 GRD composites (VV + VH), per-pixel tree-species labels (uint8), and five per-pixel tree-structure targets (mean height, median height, height variance, tree count, tree density) plus tree-count variance—all derived from the Bavarian individual-tree inventory. An optional co-registered DOP20 aerial orthophoto (20 cm GSD, 6400×6400×3 uint8) is present for patches where dop20_available=True. Data is distributed as WebDataset .tar shards plus slim Parquet index files; tensor layouts are fully described in data/schema.json.", |
| "url": "https://huggingface.co/datasets/mammmarahmed/TreeUQ", |
| "version": "1.0.0", |
| "datePublished": "2026-05-08", |
| "license": "https://creativecommons.org/licenses/by/4.0/", |
| "keywords": [ |
| "earth-observation", |
| "remote-sensing", |
| "sentinel-2", |
| "sentinel-1", |
| "aerial-imagery", |
| "DOP20", |
| "orthophoto", |
| "tree-species", |
| "tree-height", |
| "tree-inventory", |
| "uncertainty-quantification", |
| "Bavaria", |
| "Germany", |
| "geospatial", |
| "webdataset", |
| "benchmark" |
| ], |
| "citeAs": "Ahmed, A. et al. (2026). TreeUQ: A Large-Scale Earth Observation Benchmark for Tree Species and Height Estimation with Uncertainty Quantification. NeurIPS 2026 Datasets and Benchmarks Track. https://huggingface.co/datasets/mammmarahmed/TreeUQ", |
| "creator": [ |
| { |
| "@type": "Organization", |
| "name": "Bavaria EO Benchmark Authors" |
| } |
| ], |
| "distribution": [ |
| { |
| "@type": "cr:FileObject", |
| "@id": "schema-json", |
| "name": "schema.json", |
| "description": "Machine-readable tensor schema: maps each .tar member suffix to dtype, shape, band names, units, and description.", |
| "contentUrl": "data/schema.json", |
| "encodingFormat": "application/json", |
| "sha256": "43874ca9cb6ef685987d9abc3e17c520ed06c3d3ac4e08728a426476ddf65ed2" |
| }, |
| { |
| "@type": "cr:FileObject", |
| "@id": "index-train", |
| "name": "train.parquet", |
| "description": "Slim patch index for the train split (31,092 patches). Columns: patch metadata + sample_key + shard_relpath.", |
| "contentUrl": "data/index/train.parquet", |
| "encodingFormat": "application/x-parquet", |
| "sha256": "d6ee8ddf35079ef8f6ff6d386b10046cd8dffe58fa179aa6195ea1daa130f5a8" |
| }, |
| { |
| "@type": "cr:FileObject", |
| "@id": "index-validation", |
| "name": "validation.parquet", |
| "description": "Slim patch index for the validation split (8,150 patches).", |
| "contentUrl": "data/index/validation.parquet", |
| "encodingFormat": "application/x-parquet", |
| "sha256": "035b63d9b4de518d19b994df3ed683181f1affeffca5f856c2c7115083932acc" |
| }, |
| { |
| "@type": "cr:FileObject", |
| "@id": "index-test", |
| "name": "test.parquet", |
| "description": "Slim patch index for the test split (6,177 patches).", |
| "contentUrl": "data/index/test.parquet", |
| "encodingFormat": "application/x-parquet", |
| "sha256": "d1beef454be539eda368a2aea8e91915e073d08777c88e2a3e4a4301fb39dab6" |
| }, |
| { |
| "@type": "cr:FileSet", |
| "@id": "train-shards", |
| "name": "train shards", |
| "description": "WebDataset .tar shards for the train split (~1 GB each). Each sample key is a 6-digit zero-padded patch_id. Tensor members are documented in schema.json.", |
| "encodingFormat": "application/x-tar", |
| "includes": "data/shards/train-*.tar" |
| }, |
| { |
| "@type": "cr:FileSet", |
| "@id": "validation-shards", |
| "name": "validation shards", |
| "description": "WebDataset .tar shards for the validation split.", |
| "encodingFormat": "application/x-tar", |
| "includes": "data/shards/validation-*.tar" |
| }, |
| { |
| "@type": "cr:FileSet", |
| "@id": "test-shards", |
| "name": "test shards", |
| "description": "WebDataset .tar shards for the test split.", |
| "encodingFormat": "application/x-tar", |
| "includes": "data/shards/test-*.tar" |
| } |
| ], |
| "recordSet": [ |
| { |
| "@type": "cr:RecordSet", |
| "@id": "patches", |
| "name": "patches", |
| "description": "One record per patch. Metadata columns come from the Parquet index files; the split column identifies train/val/test membership. Tensor data is not stored in the Parquet rows—decode binary members from the shard using sample_key per data/schema.json.", |
| "key": { |
| "@id": "patches/patch_id" |
| }, |
| "field": [ |
| { |
| "@type": "cr:Field", |
| "@id": "patches/patch_id", |
| "name": "patch_id", |
| "description": "Globally unique integer patch identifier from the full-Bavaria grid.", |
| "dataType": "sc:Integer", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "patch_id" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/center_x", |
| "name": "center_x", |
| "description": "Patch centre easting in EPSG:25832 (metres).", |
| "dataType": "sc:Float", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "center_x" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/center_y", |
| "name": "center_y", |
| "description": "Patch centre northing in EPSG:25832 (metres).", |
| "dataType": "sc:Float", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "center_y" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/row_start", |
| "name": "row_start", |
| "description": "Top row index of the patch in the master grid (pixels).", |
| "dataType": "sc:Integer", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "row_start" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/row_end", |
| "name": "row_end", |
| "description": "Exclusive bottom row index of the patch in the master grid.", |
| "dataType": "sc:Integer", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "row_end" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/col_start", |
| "name": "col_start", |
| "description": "Left column index of the patch in the master grid (pixels).", |
| "dataType": "sc:Integer", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "col_start" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/col_end", |
| "name": "col_end", |
| "description": "Exclusive right column index of the patch in the master grid.", |
| "dataType": "sc:Integer", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "col_end" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/valid_pixel_pct", |
| "name": "valid_pixel_pct", |
| "description": "Fraction of non-NaN pixels in the Sentinel-2 bands for this patch [0, 1].", |
| "dataType": "sc:Float", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "valid_pixel_pct" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/tree_pixel_pct", |
| "name": "tree_pixel_pct", |
| "description": "Fraction of pixels with at least one inventory tree [0, 1].", |
| "dataType": "sc:Float", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "tree_pixel_pct" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/mean_tree_count", |
| "name": "mean_tree_count", |
| "description": "Mean number of inventory trees per non-zero pixel in this patch.", |
| "dataType": "sc:Float", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "mean_tree_count" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/mean_tree_count_variance", |
| "name": "mean_tree_count_variance", |
| "description": "Mean tree-count variance statistic for this patch.", |
| "dataType": "sc:Float", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "mean_tree_count_variance" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/split", |
| "name": "split", |
| "description": "Dataset split: 'train', 'val', or 'test'.", |
| "dataType": "sc:Text", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "split" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/block_col", |
| "name": "block_col", |
| "description": "Block column index for spatial train/val/test assignment.", |
| "dataType": "sc:Integer", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "block_col" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/block_row", |
| "name": "block_row", |
| "description": "Block row index for spatial train/val/test assignment.", |
| "dataType": "sc:Integer", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "block_row" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/block_id", |
| "name": "block_id", |
| "description": "Geographic block ID used for spatial train/val/test assignment.", |
| "dataType": "sc:Integer", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "block_id" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/distance_to_nearest_test_km", |
| "name": "distance_to_nearest_test_km", |
| "description": "Geodesic distance (km) from this patch centre to the nearest test-split patch centre.", |
| "dataType": "sc:Float", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "distance_to_nearest_test_km" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/buffered", |
| "name": "buffered", |
| "description": "True if the patch falls within the 5 km buffer zone around test blocks.", |
| "dataType": "sc:Boolean", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "buffered" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/in_bavaria", |
| "name": "in_bavaria", |
| "description": "True if the patch centre falls within the official Bavaria administrative boundary.", |
| "dataType": "sc:Boolean", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "in_bavaria" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/dop20_available", |
| "name": "dop20_available", |
| "description": "True if a DOP20 aerial orthophoto tile is included in the shard for this patch.", |
| "dataType": "sc:Boolean", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "dop20_available" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/sample_key", |
| "name": "sample_key", |
| "description": "WebDataset sample key: 6-digit zero-padded decimal used to address the patch inside its .tar shard (e.g. '000410').", |
| "dataType": "sc:Text", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "sample_key" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/shard_relpath", |
| "name": "shard_relpath", |
| "description": "Repo-relative path to the WebDataset .tar shard containing this patch (e.g. 'data/shards/train-000000.tar').", |
| "dataType": "sc:Text", |
| "source": { |
| "fileObject": { "@id": "index-train" }, |
| "extract": { "column": "shard_relpath" } |
| } |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/s2_spring", |
| "name": "s2_spring", |
| "description": "Sentinel-2 L2A spring median composite. 10 bands (B2 B3 B4 B8 B5 B6 B7 B8A B11 B12). DN units; divide by 10000 for reflectance. Stored as {sample_key}.s2_spring.f32 (raw float32 LE, shape 128×128×10).", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128,10" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/s2_summer", |
| "name": "s2_summer", |
| "description": "Sentinel-2 L2A summer median composite. Same band order and units as s2_spring. Stored as {sample_key}.s2_summer.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128,10" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/s2_autumn", |
| "name": "s2_autumn", |
| "description": "Sentinel-2 L2A autumn median composite. Stored as {sample_key}.s2_autumn.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128,10" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/s2_winter", |
| "name": "s2_winter", |
| "description": "Sentinel-2 L2A winter median composite. Stored as {sample_key}.s2_winter.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128,10" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/s1_spring", |
| "name": "s1_spring", |
| "description": "Sentinel-1 GRD spring composite. 2 bands (VV, VH), linear gamma-0; typical range 0–1. Stored as {sample_key}.s1_spring.f32 (shape 128×128×2).", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128,2" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/s1_summer", |
| "name": "s1_summer", |
| "description": "Sentinel-1 GRD summer composite. Stored as {sample_key}.s1_summer.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128,2" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/s1_autumn", |
| "name": "s1_autumn", |
| "description": "Sentinel-1 GRD autumn composite. Stored as {sample_key}.s1_autumn.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128,2" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/s1_winter", |
| "name": "s1_winter", |
| "description": "Sentinel-1 GRD winter composite. Stored as {sample_key}.s1_winter.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128,2" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/tree_species", |
| "name": "tree_species", |
| "description": "Tree species classification per 10 m pixel. Class ID 0 = background / no data. Stored as {sample_key}.tree_species.u8 (uint8, shape 128×128).", |
| "dataType": "sc:Integer", |
| "isArray": true, |
| "arrayShape": "128,128" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/mean_height_arr", |
| "name": "mean_height_arr", |
| "description": "Mean tree height per 10 m pixel (m); NaN = no inventory trees. Stored as {sample_key}.mean_height.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/median_height_arr", |
| "name": "median_height_arr", |
| "description": "Median tree height per 10 m pixel (m). Stored as {sample_key}.median_height.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/height_variance_arr", |
| "name": "height_variance_arr", |
| "description": "Variance of tree height per 10 m pixel. Stored as {sample_key}.height_variance.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/tree_count_arr", |
| "name": "tree_count_arr", |
| "description": "Number of inventory trees per 10 m pixel. Stored as {sample_key}.tree_count.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/tree_density_arr", |
| "name": "tree_density_arr", |
| "description": "Tree density: sum of tree_count in the 3×3 pixel neighbourhood. Stored as {sample_key}.tree_density.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/tree_count_variance_arr", |
| "name": "tree_count_variance_arr", |
| "description": "Spatial variance of tree count per 10 m pixel. Stored as {sample_key}.tree_count_variance.f32.", |
| "dataType": "sc:Float", |
| "isArray": true, |
| "arrayShape": "128,128" |
| }, |
| { |
| "@type": "cr:Field", |
| "@id": "patches/dop20_rgb", |
| "name": "dop20_rgb", |
| "description": "DOP20 true-colour aerial orthophoto at 20 cm GSD, co-registered with the 10 m Sentinel stack. Present only when dop20_available=True. Stored as {sample_key}.dop20_rgb.u8 (raw uint8 LE, shape 6400×6400×3, channels R G B).", |
| "dataType": "sc:Integer", |
| "isArray": true, |
| "arrayShape": "6400,6400,3" |
| } |
| ] |
| } |
| ], |
| "rai:hasSyntheticData": false, |
| "prov:wasDerivedFrom": [ |
| { |
| "@id": "https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR_HARMONIZED", |
| "prov:label": "Sentinel-2 L2A (Harmonised)" |
| }, |
| { |
| "@id": "https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S1_GRD", |
| "prov:label": "Sentinel-1 GRD" |
| }, |
| { |
| "@id": "https://geodaten.bayern.de/opengeodata/OpenDataDetail.html?pn=dop20rgb", |
| "prov:label": "DOP20 – Digitale Orthophotos 20 cm (Bavaria)" |
| }, |
| { |
| "@id": "https://geodaten.bayern.de/opengeodata/OpenDataDetail.html?pn=einzelbaeume", |
| "prov:label": "Einzelbäume – Bavarian Individual Tree Inventory" |
| } |
| ], |
| "prov:wasGeneratedBy": [ |
| { |
| "@type": "prov:Activity", |
| "prov:type": { |
| "@id": "https://www.wikidata.org/wiki/Q5227332" |
| }, |
| "sc:description": "Sentinel-2 seasonal median composites (spring, summer, autumn, winter 2025) were computed over Bavaria to remove cloud contamination; Sentinel-2 20 m bands (B5, B6, B7, B8A, B11, B12) were resampled to 10 m via bilinear interpolation. Sentinel-1 GRD backscatter (VV, VH) was processed to linear gamma-0 and composited per season. For each 10 m × 10 m pixel, mean, median, variance of tree height and tree count were aggregated from the Bavarian individual-tree inventory (Einzelbäume). DOP20 tiles were co-registered to the 10 m Sentinel grid (6400×6400 px per patch at 20 cm GSD). Patches were spatially blocked and assigned to train / validation / test splits with a strict 5 km buffer around test blocks to prevent spatial leakage." |
| } |
| ], |
| "rai:dataLimitations": "The dataset has several known constraints and domain restrictions. Geographically, the dataset is restricted to the federal state of Bavaria, meaning models trained on this data may face distributional gaps and may not generalize to other global regions without domain adaptation. Targets are provided at 10 m resolution, so individual trees cannot be directly observed and their properties must be inferred from medium-resolution radiometric and structural signals. Ground truth labels are sparse and unevenly distributed—a median valid label coverage of approximately 26% per patch—which may bias learning and skew model behaviour toward managed landscapes compared to dense natural forests. Seasonal composites remove short-term dynamics, which may obscure fine-grained temporal variability.", |
| "rai:dataBiases": "Significant selection and label bias exists due to extreme label sparsity. Only a small fraction of grid cells contain tree observations. This highly imbalanced supervision may bias learning, potentially causing models to perform better on managed, urban, or roadside trees compared to dense natural forest interiors. The Bavarian individual-tree inventory itself reflects cadastral and forestry-management priorities and does not uniformly sample the ecologically full tree population.", |
| "rai:personalSensitiveInformation": "The dataset consists entirely of satellite imagery, aerial orthophotos, and forest structure measurements. It contains no personal or sensitive human information. All spatial data is aligned to a 10 m reference grid (EPSG:25832) covering Bavaria; the dataset does not include location data at a resolution that could identify individuals.", |
| "rai:dataUseCases": "TreeUQ is intended to measure fundamental, discrete ecological quantities of forest structure—specifically tree density and average tree height—along with their associated variances for uncertainty quantification. Validated use cases include: multi-task learning (joint prediction of continuous height and discrete count from shared satellite inputs), cross-modal knowledge distillation (using 20 cm high-resolution DOP20 RGB together with 10 m Sentinel data), spatial generalization benchmarking, and the evaluation of Earth Observation foundation models. Supporting baselines using XGBoost, U-Net, SegFormer-B3, and the Clay Foundation Model are provided.", |
| "rai:dataSocialImpact": "Positive Effects: By enabling accurate large-scale characterisation of forest structure and integrating uncertainty quantification, TreeUQ can support downstream decision-making in climate science, carbon cycle modelling, ecosystem resilience assessment, biodiversity monitoring, and sustainable forest management. Negative Effects and Fairness: A potential risk is the misuse of models trained on this benchmark for automated environmental policy-making or carbon accounting in regions dissimilar to Bavaria or in unmanaged natural forests. Because ground truth over-represents managed trees and is strictly localised, applying these models elsewhere could yield unfair, biased, or highly inaccurate ecological assessments. Mitigations: The dataset is defined strictly for methodological research rather than operational deployment. A rigorous deterministic spatial block-splitting strategy with a 5 km buffer zone is enforced to prevent geographic data leakage. A Masked Smooth L1 loss is recommended in evaluation protocols to handle label sparsity fairly." |
| } |
|
|
