Directional distributions for multi-dimensional random point processes (Q4031396)

For investigating simultaneously the orientation and distance aspects of anisotropic point processes two types of directional distributions are considered. Both of them are based on counting procedures of suitable point sets. The first one is connected with the direction of the nearest neighbour within the set of those points, whose distance to the given point is not smaller than some lower bound \(r>0\). The second one results from counting the numbers of points in sectors of some given spherical shell with inner radius \(r\) and outer radius \(r'\) centered at a point of the point process. Both types of directional distributions are defined for not necessarily stationary point processes and are characterized by means of conditional distributions. It is shown, how these quantities can be simplified assuming stationarity respectively stationarity and isotropy. Finally some estimators of these directional distributions for stationary point processes are discussed.