Hi all,
Here is an interesting problem that I could use some suggestions and/or
sample code for. I have a street network where all the street segments
are noded at intersections. What I need to extract is all the blocks in
the coverage. A "block" is the smallest polygon bounded by segments that
does not have any segment crossing it. It could have a cul-de-sac
protruding into its interior as long as it does not divide the polygon
in two.
I could create a node adjacency list, but I'm not sure if that helps me.
I'm thinking that each edge needs to be tagged with a right and left
polygon id, then you can create polygons by grouping on the id.
while (current = get_edge_without_rl_pid) {
if current.right_side_edge has no pid {
get new_pid as this_pid
get other edges from end
find right most edge
if edge has id then {
update all new_pid to id
mark new_pid as unused
this_pid = id
}
else
mark both edges as this_pid
}
if current.left_side_edge has no pid {
get new_pid as this_pid
get other edges from end
find left most edge
if edge has id then {
update all new_pid to id
mark new_pid as unused
this_pid = id
}
else
mark both edges as this_pid
}
}
Does this seem like it would work?
Anyone interested in turning this into plpgsql?
It might be better to recursively pick an edge and recursively walk
connected right most edges back onto itself marking the edges as you go.
At some point you may still have edges not included in any of the above
that can only be marked by walking the left most edges.
Thoughts?
-Steve W
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
