The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Dataset for the repository "MoE-DSI-3D".
Motivation
To evaluate the applicability of DSI-3D and MoE-DSI-3D on a large-scale dataset, we use the LiDAR HD (LiDAR Haute Densité) dataset. LiDAR HD covers the French territory and provides an average point density of approximately 10 points/m².
The dataset was acquired through a country-wide, tile-based LiDAR campaign covering a wide variety of environments. We selected the Paris area as our primary study region because of its diverse urban environments.
We focus on two areas:
Paris_East: a 9 km² area extending from Saint-Mandé to Vincennes. Paris_West: a 16 km² area covering parts of the Bois de Boulogne and extending to the Champ-de-Mars.
Construction
To construct our dataset, we subdivided the LiDAR HD tiles into regular grids:
Paris_East: 150 × 150 m cells. Paris_West: 200 × 200 m cells.
To provide overlap between neighboring scenes, each core cell at position (x,y) is expanded by aggregating its neighboring cells within a range of ±2 cells along both axes. This results in an input scene covering 100 × 100 m, composed of 25 neighboring sub-cells.
In other words, each 20 × 20 m core region is embedded within a larger 100 × 100 m context window, providing the surrounding spatial information required by DSI-3D and MoE-DSI-3D.
- Downloads last month
- 61