Even if the detection problem is not given, the interaction system
can find a previously undetected building, just by giving the information that
for sure there is a building in the area.
Another example of a detection problem would be that a part of an L-
or T-shaped building is not detected. After the problem is specified,
this can be corrected by clicking on the roof of the missing part.
There are two ways of correcting a hypothesis generated by the initial
step (see also Corrective Interaction):
here the second alternative is presented: one can choose to extract
edges and corners and associate them to a part of the building model.
In this example, a roof-side of the building is specified by an edge
extracted in the image. Then this edge is added to the current
hypothesis.
The last example shows that the manual feature extraction and the following fitting and
verification steps can also be applied to buildings that are
automatically detected, but partially wrong.
first building
second building