/ Filters / Chapter 35. Using spatial properties of features in filters / Numbers of feature intersections and inclusionsThis property means that the tested feature is fully located within the argument feature.
The tested feature can belong to any spatial category. Argument features are to belong to one of categories which have inner region, that is, they can be area, text or image feature.
For an area (or image, or text) tested feature it is possible that its boundary coincides partially or entirely with boundary of an argument feature. Full coincidence of features is also considered as their inclusion.
Examples are given on the figure 35-7.
a) There is inclusion
b) No inclusion
Figure 35-7. Inclusion property for area features
On the figure 35-7, a tested features T1 and T2 are included into argument features A1 and A2, correspondingly. On the figure 35-7, b different cases are shown of such relative placement where tested features T1 and T2 are not included into argument features A1 and A2.
For a line tested feature it is required that all its polyline segments lie within inner region of the argument, but polyline vertices can lie on the boundary as well. Examples are given on figures 35-8, a, b.
a) There is inclusion
b) No inclusion
Figure 35-8. Inclusion property for line features
For a point tested feature inclusion occur only if this feature (to be more exact, its base point) lies within inner region of the argument feature but not on its boundary (figure 35-9).
a) There is inclusion b) No inclusion
Figure 35-9. Inclusion property for point features
It is to be noted that all coordinate calculations concerned with determining of features relative placement are performed with a small inaccuracy which value depends on the preciseness of representing real numbers in the program. Therefore checking exact coincidence of a line or a point with a boundary of an area feature is not quite reliable. For instance, a point which lies exactly on a feature's boundary can be classified as lying within or outside the feature owing to the calculation inaccuracy.
So, checking the Inclusion property can be considered as reliable only in the case when the tested feature lies completely within an inner region of an argument feature and does not touch its boundaries.
It is not difficult to see that the number of inclusions of a tested feature into argument features can be greater than 1 only if the argument features themselves intersect each other, as it is shown on the figure 35-10.
Figure 35-10. Inclusion into two argument features
 |  |
