People who engage in photogrammetry have a problem. In order to determine the relative dimensions of a thing, the size of a thing, you need to have a known set of distances in a number of the shots to provide relational information about what measurements go where.
This is usually done by putting flat paper targets down around the subject, but flat targets have to be measured in relation to each other. Flat objects can only provide dimensional information in two dimensions, sensibly enough. Getting three-dimensional measurements can be challenging.
Enter: the photogrammetric target cube. The target cube is of a known size, 4 cm on a side. Three of the edges are orthogonal and edged while the other three are beveled, making them visually distinct. The cube itself is visually distinct no matter which direction you look at it from, making it easier for structure from motion software like VisualSFM, Agisoft Photoscan, or Autodesk ReMake to derive its spatial relationships.