As an example, consider inspection of an object by a stereo head. We define the head position as the midpoint of the line connecting the optical centres of the cameras.

The stereo visibility algorithm takes as input a portion of the property sphere describing a visibility region (of a feature or set thereof), in which every viewpoint is weighted by an optimality coefficient encoding both visibility and reliability of inspection (notice that such a region may not be maximally connected).

The output of the algorithm is an ordered list of head positions ( c₁ ,.....,c_k ), c_i = ( V_il , V_ir ) , where ( V_il , V_ir ) , are the viewpoints of the left and right cameras. The global complexity of the stereo positioning algorithm is R(O (F) + O(RC)) = O(RF) + O(R₂ C), where R is the number of viewpoints in the visibility region, C is the number of viewpoints at a fixed distance from a given one (a function of the dome's resolution), and F is the number of viewpoints on the geodesic dome. If O(F) < O(RC), the complexity becomes O(R₂ C). Figure 15, below, shows (shaded) the optimum position for a stereo head of baseline 400mm to inspect the planar face on the bottom side of the object, nearest the viewer, having the inverted `V' bounding contour. The focal length of the cameras was 20mm, the dome radius 1376mm. The stand is about 16cm tall. This solution was found in about two seconds.