{"success":true,"data":{"author":"MPG Ranch (Kyle Doherty, Erik Samose, Max Gurinas, Brandon Trabucco, Ruslan Salakhutdinov)","bucketAccessType":"public","bucketName":"mpg-ranch/drone-lsr","bucketPrefix":"","category":"Earth Observation & Geospatial","cloudProvider":"huggingface","createdAt":"2026-11-19T15:09:27.848Z","description":"Aerial RGB orthomosaic tiles of the same ground captured at three times of day (10:00, 12:00 and 15:00) on 7 November 2024 over Lower Partridge Alley at MPG Ranch, Montana, with a co-registered canopy height model. 620 tile sets of 1024 x 1024 pixels at 1.2 cm ground resolution; canopy height is an int32 grid in centimeters. Three configurations share tile ids and an 80/20 train/test split: default (the three images plus canopy height), dinov2_base (DINOv2 ViT-B/14 CLS and patch tokens per image) and dinov3_sat (DINOv3 ViT-L/16 satellite-pretrained, SAT-493M, CLS and patch tokens). Tiles with transient objects such as vehicles were removed. Built to train and test vision encoders whose features stay stable under changing illumination and shadows, and to compare natural-image and satellite-pretrained encoders for drone remote sensing of vegetation. Parquet, about 6.7 GB; load with datasets.load_dataset('mpg-ranch/drone-lsr', config). Licensed CC BY 4.0; cite Doherty, Samose, Gurinas, Trabucco and Salakhutdinov (MPG Ranch, 2024).","displayTitle":"Light Stable Representations (LSR): Drone Orthomosaic Tiles at Three Times of Day with Canopy Height, MPG Ranch, Montana","huggingFaceUrl":"https://huggingface.co/datasets/mpg-ranch/drone-lsr","id":"a1962337-c1b9-544c-a914-a0a6d02050d0","license":"CC-BY-4.0","licenseEvidenceUrl":"https://huggingface.co/datasets/mpg-ranch/drone-lsr","objectCountEstimate":24,"probeStatus":"ok","referenceUrl":"https://huggingface.co/datasets/mpg-ranch/drone-lsr","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"mpg-ranch/drone-lsr\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"mpg-ranch/drone-lsr\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=mpg-ranch, license:cc-by-4.0, size_categories:1K<n<10K, format:parquet, downloads=401"],"format":"mixed","generatedAt":"2026-11-19T15:21:54.781Z","layout":{"example":"data/test-00000-of-00003.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"50415231150415aa0115724c1514150012000055200500000031305f3131110900351109043904091b00320d0800350d08003605210c315f313011090031051a28315f320400000031315f391500152015242c15141510150615060000103c020000001401040510325476980000001504158cefe90115c4f0e9014c15021500120000c6f7745cc23b1d0089504e470d0a1a0a0000000d49484452000004000104f4e3ff08060000007f1d2b830001000049444154789c64fd6793244976250a1e5553e36e4ec38325adacae261834fa0d46f67dddf79776f787eddb95fd3a32789019000d0c1a8d26555ddd459305756a9ca9ee87ab573d12539014546765467898a95e72ee39e78affc7ffebff69eabac6300c1886015555a1284a34750b630c3ccf8310025a6b8c23fdeaba16755da1ef7b4829e0073e269314f3f90cabd5128bc502abd50a1717e7984e33b44d83cd6683c7c747dcdddde1f17183222fd1753d00812008e0791edab6435d3730c6200802f8be6f3f5389b66d20a584effb909e443ff4689a067d3f40080129a5fbe5791e7cdf47100408c31042080cc380711c318e230001a3816118108621e23881d606799ea3aa2a68ad1186e1277fd71803dff7118621a238429aa50080fd7e8fc36e0f2904a22882e77930c6601c47f47d8fb66d51d535eab6c1309c3e2b7f46dff7618c0100c4718ce9748a288ad0","encoding":"hex","key":"data/test-00000-of-00003.parquet","truncatedAt":512}],"sampleObjects":["data/test-00000-of-00003.parquet","data/test-00001-of-00003.parquet","data/test-00002-of-00003.parquet","data/train-00000-of-00010.parquet","data/train-00001-of-00010.parquet"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\ntask_categories:\n- feature-extraction\n- image-to-image\nlanguage:\n- en\ntags:\n- remote-sensing\n- aerial-imagery\n- orthomosaic\n- lighting-invariance\n- representation-stability\n- vision-encoder\n- time-series\n- dinov2\n- dinov3\n- embeddings\n- multi-config\npretty_name: Light Stable Representations\nsize_categories:\n- n<1K\ndataset_info:\n- config_name: default\n  features:\n  - name: idx\n    dtype: string\n  - name: image_t0\n    dtype: image\n  - name: image_t1\n    dtype: image\n  - name: image_t2\n    dtype: image\n  - name: canopy_height\n    dtype:\n      array2_d:\n        shape:\n        - 1024\n        - 1024\n        dtype: int32\n  splits:\n  - name: train\n    num_bytes: 4905235380\n    num_examples: 487\n  - name: test\n    num_bytes: 1221459061\n    num_examples: 122\n  download_size: 3688072446\n  dataset_size: 6126694441\n- config_name: dinov2_base\n  features:\n  - name: idx\n    dtype: string\n  - name: cls_t0\n    list: float32\n    length: 768\n  - name: cls_t1\n    list: float32\n    length: 768\n  - name: cls_t2\n    list: float32\n    length: 768\n  - name: patch_t0\n    dtype:\n      array2_d:\n        shape:\n        - 256\n        - 768\n        dtype: float32\n  - name: patch_t1\n    dtype:\n      array2_d:\n        shape:\n        - 256\n        - 768\n        dtype: float32\n  - name: patch_t2\n    dtype:\n      array2_d:\n        shape:\n        - 256\n        - 768\n        dtype: float32\n  splits:\n  - name: train\n    num_bytes: 1154971327\n    num_examples: 487\n  - name: test\n    num_bytes: 289335733\n    num_examples: 122\n  download_size: 1487171455\n  dataset_size: 1444307060\n- config_name: dinov3_sat\n  features:\n  - name: idx\n    dtype: string\n  - name: cls_t0\n    list: float32\n    length: 1024\n  - name: cls_t1\n    list: float32\n    length: 1024\n  - name: cls_t2\n    list: float32\n    length: 1024\n  - name: patch_t0\n    dtype:\n      array2_d:\n        shape:\n        - 196\n        - 1024\n        dtype: float32\n  - name: patch_t1\n    dtype:\n      array2_d:\n        shape:\n        - 196\n        - 1024\n        dtype: float32\n  - name: patch_t2\n    dtype:\n      array2_d:\n        shape:\n        - 196\n        - 1024\n        dtype: float32\n  splits:\n  - name: train\n    num_bytes: 1180053775\n    num_examples: 487\n  - name: test\n    num_bytes: 295619221\n    num_examples: 122\n  download_size: 1520934285\n  dataset_size: 1475672996\nconfigs:\n- config_name: default\n  data_files:\n  - split: train\n    path: data/train-*\n  - split: test\n    path: data/test-*\n- config_name: dinov2_base\n  data_files:\n  - split: train\n    path: dinov2_base/train-*\n  - split: test\n    path: dinov2_base/test-*\n- config_name: dinov3_sat\n  data_files:\n  - split: train\n    path: dinov3_sat/train-*\n  - split: test\n    path: dinov3_sat/test-*\n---\n\n# Light Stable Representations Dataset\n\n## Dataset Description\n\nThis dataset contains aerial orthomosaic tiles captured at three different times of day (10:00, 12:00, and 15:00). The dataset is organized into three configurations: `default` (raw images + canopy height), `dinov2_base` (DINOv2 embeddings), and `dinov3_sat` (DINOv3 embeddings). All configurations share consistent train/test splits with matching tile identifiers for cross-referencing. The dataset is designed for training vision encoders that maintain consistent feature representations despite changes in illumination, with applications in remote sensing and environmental monitoring.\n\n## Dataset Configurations\n\nThe dataset is organized into three configurations, each serving different research needs:\n\n### Configuration: `default`\nRaw imagery and environmental data for direct analysis:\n\n| Feature | Type | Shape | Description |\n|---------|------|--------|-------------|\n| `idx` | string | - | Tile identifier in format `{ROW}_{COL}` for geographic referencing |\n| `image_t0` | Image | 1024×1024×3 | Morning capture at 10:00 AM (time=1000) |\n| `image_t1` | Image | 1024×1024×3 | Noon capture at 12:00 PM (time=1200) |\n| `image_t2` | Image | 1024×1024×3 | Afternoon capture at 3:00 PM (time=1500) |\n| `canopy_height` | int32 | [1024, 1024] | Canopy height grid in centimeters from can","encoding":"utf8","key":"README.md","name":"README.md","size":10389,"truncated":true}],"sourceObjectKey":"data/test-00000-of-00003.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"mpg-ranch/drone-lsr\")\nprint(ds)","starterSamplePath":"data/test-00000-of-00003.parquet","totalObjects":24},"summary":"Aerial RGB orthomosaic tiles of the same ground captured at three times of day (10:00, 12:00 and 15:00) on 7 November 2024 over Lower Partridge Alley at MPG Ranch, Montana, with a…","title":"Light Stable Representations (LSR): Drone Orthomosaic Tiles at Three Times of Day with Canopy Height, MPG Ranch, Montana","updatedAt":"2026-11-19T19:01:08.671Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/mpg-ranch/drone-lsr"}}