|
G'd day all,
For my mapping work I use very detailed polyline and contour
data (amongst others), which have been produced from 1:50,000 scale maps
through digitizing. The line data are extremely finely sampled and hence
reproduce beautifully all curved portions of lines at that scale. However, the
ultimate scale at which my maps will be produced is 1:200,000 or 1:400,000, a
linear reduction of 1 in 4 or even 1 in 8 and the precision of the original data
is therefore way too high for the end product.
Question is: Is anyone aware of a routine through which one can pass
the data files, and which will reduce the data density in such a way that
at the smaller scale the exactness is still maintained. There is a mathematical
rule which says that the spatial sampling frequency is dictated by the
'curviness' of the sampled line in order to restore the original curve
without aliasing. (The word Nyquist frequency comes to mind).
Anyone has any thoughts on this??
TIA
andr�
from: andr� boessenkool
po box 101 - vlottenburg 7604 south africa tel (+27 21) 881 3188 fax (+27 21) 881 3189 [EMAIL PROTECTED] |
