Tiling
Tiling options
Select how to divide the reconstruction task.
3D models generated by ContextCapture can span entire cities and hence may not fit in your computer's memory in one piece. Thus, it is often necessary to divide them into tiles.
Four tiling modes are available:
- No tiling: Do not subdivide reconstruction.
- Regular planar grid: Divide reconstruction into square tiles along the XY plane.
- Regular volumetric grid: Divide reconstruction into cubic tiles.
- Adaptive tiling: Adaptively subdivide reconstruction into boxes to meet a target memory usage. This is particularly useful to reconstruct a 3D model with a highly non uniform resolution, e.g. when reconstructing a city from aerial images and ground images of a few landmarks. In such a case, it is not possible to find a regular grid size adequate for all areas.
When using a tiling other than an adaptive tiling, select tile size and tiling origin (click on Advanced options) depending on your needs.
Tiling is aligned to the selected spatial reference system. Tiles are discarded if they are completely outside of the region of interest, or if they do not contain any block tie point.
By default, tiles not containing any tie point are discarded from the tiling, enable the option Discard empty tiles to keep all tiles in the specified region of interest.
For georeferenced reconstructions, tiling can be exported in KML format (menu Reconstruction > Tiling > Export to KML ...) to get an overview of the spatial framework in a standard GIS tool or in Google Earth.
Bing Maps Tile System
ContextCapture allows to perform reconstruction following the Bing Maps Tile System.
Select a Bing Maps spatial reference system to setup a reconstruction using a Bing Maps coordinate system and tiling for a given level.