Graphics Reference
In-Depth Information
Figure
.
.
Finding regions with vegetation using the angle query (let) between the B
andB
axes.
Note the arrowheads on the B
axis, which specify the range of angles for the selected lines
in the projective plane (the directions are the ideal points) rather than the Euclidean
plane. For sets of points that have a “general” direction with a negative slope (i.e.,
are “negatively correlated”), the lines representing them in
-coords cross each other
in-between the axes, and they can be selected with the pinch query. For positively
correlated sets of points, their corresponding lines cross outside the axes and can be
selected with the angle query. All of this exemplifies the need to understand some
of the basic geometry in order to use the queries effectively and of course to design
them properly.Now that we have introduced the three atomic queries, the next stage
is to show how they can be combined to construct complex queries.
However, prior to this, Fig.
.
(let) begs the question: “what if the B
andB
axes were not adjacent?” henthe pattern andhencetheir pairwise relation wouldbe
missed. Hence the permutation of the axes used for the exploration is important. In
particular, what is the minimum number of permutations among N-axes containing
the adjacencies forallpairsofaxes?ItturnsoutWegman(
)that M permutations
