Hilbert Curve and Morton Curve #368
Conversation
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Add Morton Curve Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
|
@dr-jts nice! You know about SFCurve as a project, right? https://github.com/locationtech/sfcurve Shout if I can ever help with space-filling curves for JTS.:) |
|
Ah right... there is that. Would be interesting to compare. My implementation is much less ambitious. The ultimate goal is to implement a Hilbert R-tree, to see how that compares to the STR-packed Rtree. And also pretty pictures... |
|
Interesting... are you looking at Hilbert R-Trees as static or dynamic data structures? (I skimmed the Wikipedia article, and I haven't looked enough to understand the benefits.) Incidentally, if the Hilbert implementation matters, @cne1x could come and share what he knows about 'compact' Hilbert curves. (They are useful if the magnitude of the various dimensions are unequal. E.g., if you have 10 bits of info for x and 6 bits for y.) |
|
@jnh5y The Javadoc for HilbertCode and MortonCode could use a review if you are so inclined... |
|
As a alternate packing method for the current static R-tree structure. (Which incidentally really needs to be refactored to support this, i.e. to separate the packing strategy from the actual R-tree structure and algorithms. I'm also going to dump the SIRtree complexity, since it's no longer used in JTS having been superceded by the |
* Add HilbertCode and HilbertCurveBuilder * Add MortonCode and MortonCurveBuilder Signed-off-by: Martin Davis <mtnclimb@gmail.com>
* Add HilbertCode and HilbertCurveBuilder * Add MortonCode and MortonCurveBuilder JTS' commit locationtech/jts@856db45
Hilbert Curve and ordering
Morton Curve and ordering