Project tree

A Regard3D project is organized as follows:

  • The first level is the project. On this level the directory is defined where all the lower level items will be stored.
  • The next level are the picture sets. Each picture set defines the set of pictures used for the subsequent steps. If a project has more than one picture sets, those picture sets are completely independent: They can contain (partially) the same pictures or completely different ones.
  • The third level are the computed matches. Starting from the picture set, the user can define parameters for computing the matches. The results of this step are stored in the "Matches" level.
  • The fourth level is the triangulation. Using the matches from the previous level a triangulation is computed.
  • The fifth level is the densification. Starting from the triangulation (sparse point cloud and camera positions) a dense point cloud is computed.
  • The last level is the surface which is computed from the point cloud.

One can think of the project tree as a directory structure: The project is the topmost directory, each level adds another directory.

The project tree is visible in the top left part of the Regard3D window. A filled project tree looks like this:


Similar to the directory tree in the Windows Explorer or OS X Finder, this tree can be expanded and collapsed by clicking on the small triangle to the left of an item. By clicking on an item, some information is displayed in the information below the tree.

Most items can be double-clicked: Triangulations, Densifications and Surfaces will be loaded and displayed in the 3D view. Picture sets and Matches are displayed in a separate window.

Actions can be executed by either right-clicking an item and choosing an action in the context menu, or by selecting an item and clicking on a button in the information area below.

Items can also be deleted. But by deleting an item, the sub-items will be deleted as well. For example, if a densification shall be deleted, all surfaces that were computed from this densification will be deleted as well.

The motivation behind this tree structure is to allow the user to easily run and compare computing steps with different parameters. For example, it can be difficult to find a good parameter for a surface generation. With this tree structure it is easy to run several surface generations with slightly different parameters and so finding the best parameter set.


Go to next article: 3D View