No Wolf in the Meadow (NWM)

Three types of archives are provided: nwm_<YYYY-MM-DD>.zip — Images and camera info ( camera_info.json ) for a single timestep. nwm_pre_trained_timestep_<YYYY-MM-DD>.zip — COLMAP model and pretrained 3DGS and PuP-3DGS models for a single timestep, used as input for ChronoFuseGS. nwm_<subset>.zip — Multi-temporal model trained and fused with ChronoFuseGS for a given subset ( complete , autumn , flooding , snow , car ).

Individual Timesteps We split the data into 11 timesteps in total. The two detail timesteps recorded at 2026-01-10 ( 2026-01-10_detail_1 , 2026-01-10_detail_2 ) are only included in the Snow subset. The list of timesteps is shown below: Timestep Number of Input Images Environmental Conditions 2025-08-25 226 Green vegetation 2025-11-11 271 Autumn foliage 2025-11-22 255 Bare trees 2025-11-23 283 Bare trees, light snow, sunset 2025-12-04_1 284 Bare trees 2025-12-04_2 198 Bare trees 2025-12-08 223 Bare trees 2026-01-10 273 Snow cover 2026-01-10_detail_1 180 Snow cover 2026-01-10_detail_2 155 Snow cover 2026-03-28 208 Early spring, bare trees, flooding Each timestep's image set is split into training and test images.

Every 8th image is held out as a test image; the remaining images are used for training. The type field in camera_info.json indicates the split for each image ( "train" or "test" ). Multi-Temporal Models We grouped these individual timesteps into the multi-temporal datasets of NWM, including the Complete dataset which covers all recording days, and smaller focused subsets.

The table includes the number of registered images in the combined multi-temporal models. Subset Num T. Days Timesteps Registered Images Complete 9 8 2025-08-25 , 2025-11-11 , 2025-11-22 , 2025-11-23 , 2025-12-04_1 , 2025-12-04_2 , 2025-12-08 , 2026-01-10 , 2026-03-28 2,221 Autumn 4 4 2025-08-25 , 2025-11-11 , 2025-11-22 , 2025-12-08 975 Flooding 3 3 2025-12-04_2 , 2025-12-08 , 2026-03-28 629 Snow 3 1 2026-01-10 , 2026-01-10_detail_1 , 2026-01-10_detail_2 608 Car 3 1 2025-12-04_1 , 2025-12-04_2 , 2025-12-08 705 File Structure nwm_<YYYY-MM-DD>.zip — Contains the extracted drone video frames and camera metadata for a single recording day. nwm_<YYYY-MM-DD>/ └── data/ ├── input/ # extracted drone video frames │ ├── <image>.png │ └── ... └── camera_info.json nwm_pre_trained_timestep_<YYYY-MM-DD>.zip — Extends the individual timestep archive with a COLMAP sparse reconstruction, a pretrained 3DGS model, and PuP-3DGS compressed models. nwm_pre_trained_timestep_<YYYY-MM-DD>/ ├── data/ │ ├── images/ # images registered by COLMAP │ ├── input/ # original extracted frames │ ├── sparse/ # COLMAP sparse reconstruction │ └── camera_info.json ├── output/ │ └── point_cloud/ │ ├── init_0/ # initial 3DGS model │ └── iteration_30000/ # model after 30,000 refinement iterations │ ├── point_cloud.ply │ ├── activation.ply │ └── camera.json └── pup/ # PuP-3DGS compressed models nwm_<subset>.zip — Contains the ChronoFuseGS multi-temporal model trained and fused for a given subset.

The structure mirrors the pretrained timestep archives without the pup/ folder. nwm_<subset>/ ├── data/ │ ├── images/ │ ├── input/ │ ├── sparse/ │ └── camera_info.json └── output/ └── point_cloud/ ├── init_0/ └── iteration_30000/ ├── point_cloud.ply ├── activation.ply └── camera.json The key output files are: point_cloud.ply — 3D Gaussian attributes (position, color, covariance, opacity, etc.) without any time-specific information. activation.ply — Per-Gaussian, timestep-dependent data: the light compensation vector and opacity manipulation vector. camera.json — Camera parameters for the trained model. camera_info.json — Per-image metadata including GPS coordinates, altitude, and train/test split.

See the Camera Positions section for a full description. Camera Positions We extracted GPS coordinates from the camera logs and provide per-image coordinate information in the corresponding camera_info.json file. [ { "image" : " 006_DJI_20250829153742_0593_D_1.png " , "video_frame" : 60 , "src_video" : " 006_DJI_20250829153742_0593_D.MP4 " , "latitude" : 48.206373 , "longitude" : 16.230399 , "rel_alt" : 20.0 , "abs_alt" : 317.406 , "ecef" : [ 4089114.54278001 , 1190350.6104078596 , 4732436.830631929 ], "registered" : true , "type" : " test " , "t" : 0 }, { "image" : " 006_DJI_20250829153742_0593_D_2.png " , "video_frame" : 180 , "src_video" : " 006_DJI_20250829153742_0593_D.MP4 " , "latitude" : 48.206436 , "longitude" : 16.2306 , "rel_alt" : 20.0 , "abs_alt" : 317.406 , "ecef" : [ 4089105.352015008 , 1190363.495616108 , 4732441.499502797 ], "registered" : true , "type" : " train " , "t" : 0 }, ... ]