Thank you for all the incredibly valuable and helpful replies.
Summarizing the responses, it seems that kriging on river networks as well
as the group in Tokyo that is developing SANET are the closest exiting
projects on this topic.
Markus
[[alternative HTML version deleted]]
__
Hello.
Following papaer may be helpful.
OKABE, A., SATOH, T. and SUGIHARA, K. "A KERNEL DENSITY
ESTIMATION METHOD FOR NETWORKS,
ITS COMPUTATIONAL METHOD, AND A GIS-BASED TOOL",
Discussion Paper, No. 80, Center for Spatial Information
Science, Univ. of Tokyo,
http://www.csis.u-tokyo.ac.jp/dp/8
I don't know about road networks, but I know of at least four groups
that have worked on kriging over river networks, and one or two that
worked on coastal data using water distance instead of Euclidian
distance. Googling "kriging river network" gave quite a few hits.
--
Edzer
Markus Loecher w
Dear geo experts,
has anyone looked into kriging of spatial processes that do not live in a 2D
continuum but instead are constrained to a network/graph (e.g. a street
grid) ?
Clearly, distances need to be redefined but more than that, the covariance
matrix is a very different animal.
Thanks!
Mark
