RST Spline Surfaces - Revision history

Neteler: +https://www.youtube.com/watch?v=eLYsx-ADOGI spline interpolation lecture

2017-06-30T05:50:07Z

+https://www.youtube.com/watch?v=eLYsx-ADOGI spline interpolation lecture

← Older revision		Revision as of 05:50, 30 June 2017
Line 19:		Line 19:

	* See the module help pages		* See the module help pages
			** [[https://www.youtube.com/watch?v=eLYsx-ADOGI Video lecture on spline interpolation]] by Prof Helena Mitasova
	: '''''(TODO: write user guide)'''''		: '''''(TODO: write user guide)'''''

Neteler: /* ToDo */

2010-11-08T07:54:20Z

ToDo

← Older revision		Revision as of 07:54, 8 November 2010
Line 96:		Line 96:

	[[Category:FAQ]]		[[Category:FAQ]]
			[[Category:Interpolation]]
	[[Category: Parallelization]]		[[Category: Parallelization]]

⚠️HamishBowman: /* Tuning parameters and Validation */ hint from John

2010-02-28T19:22:01Z

Tuning parameters and Validation: hint from John

← Older revision		Revision as of 19:22, 28 February 2010
Line 27:		Line 27:
	:: See C-V script: http://www.grassbook.org/examples_menu3rd.php (Script to perform cross validation in GRASS/RST splines interpolation) for parametrisation		:: See C-V script: http://www.grassbook.org/examples_menu3rd.php (Script to perform cross validation in GRASS/RST splines interpolation) for parametrisation
	:: See this [http://lists.osgeo.org/pipermail/grass-user/2007-May/039530.html post to the mailing list]		:: See this [http://lists.osgeo.org/pipermail/grass-user/2007-May/039530.html post to the mailing list]

			* To help speed things up during tuning tests use a smaller area, e.g. 100x100 cells, that is representative of the terrain.

	=== Getting rid of those segmentation boxes ===		=== Getting rid of those segmentation boxes ===

⚠️HamishBowman: /* Tuning parameters and Validation */

2010-02-28T19:17:12Z

Tuning parameters and Validation

← Older revision		Revision as of 19:17, 28 February 2010
Line 26:		Line 26:
	:: See manual page of {{cmd\|v.surf.rst}}		:: See manual page of {{cmd\|v.surf.rst}}
	:: See C-V script: http://www.grassbook.org/examples_menu3rd.php (Script to perform cross validation in GRASS/RST splines interpolation) for parametrisation		:: See C-V script: http://www.grassbook.org/examples_menu3rd.php (Script to perform cross validation in GRASS/RST splines interpolation) for parametrisation
	:: See this [http://lists.osgeo.org/pipermail/grass-user/2007-May/039530.html to the mailing list]		:: See this [http://lists.osgeo.org/pipermail/grass-user/2007-May/039530.html post to the mailing list]

	=== Getting rid of those segmentation boxes ===		=== Getting rid of those segmentation boxes ===

⚠️HamishBowman: +link to tuning thread

2010-02-28T19:16:40Z

+link to tuning thread

← Older revision		Revision as of 19:16, 28 February 2010
Line 16:		Line 16:
	* {{cmd\|r.fillnulls}}		* {{cmd\|r.fillnulls}}

	=== Tuning parameters ===		=== Tuning parameters and Validation ===

	* See the module help pages		* See the module help pages
	: '''''(TODO: write user guide)'''''		: '''''(TODO: write user guide)'''''

	~~=== Validation ===~~
	* Cross-validation		* Cross-validation
	: '''''(TODO: write How-to)'''''		: '''''(TODO: write How-to)'''''

	:: See manual page of v.surf.rst		:: See manual page of {{cmd\|v.surf.rst}}
	:: See C-V script: http://www.grassbook.org/examples_menu3rd.php (Script to perform cross validation in GRASS/RST splines interpolation) for parametrisation		:: See C-V script: http://www.grassbook.org/examples_menu3rd.php (Script to perform cross validation in GRASS/RST splines interpolation) for parametrisation
			:: See this [http://lists.osgeo.org/pipermail/grass-user/2007-May/039530.html to the mailing list]

	=== Getting rid of those segmentation boxes ===		=== Getting rid of those segmentation boxes ===
Line 47:		Line 46:
	ERROR: input failed!		ERROR: input failed!

	In the v.surf.rst/v.vol.rst source code, there is a MAXPOINTS value defined in the <tt>surf.h</tt> file. MAXPOINTS simply says how many input points are allowed to be computed without segmentation. I had to increase the value of MAXPOINTS in <tt>surf.h</tt> to 1346 (or a bit more), recompile, and then use segmax=1346 to avoid segmentation.		In the {{cmd\|v.surf.rst}}/{{cmd\|v.vol.rst}} source code, there is a MAXPOINTS value defined in the <tt>surf.h</tt> file. MAXPOINTS simply says how many input points are allowed to be computed without segmentation. I had to increase the value of MAXPOINTS in <tt>surf.h</tt> to 1346 (or a bit more), recompile, and then use segmax=1346 to avoid segmentation.
	Tricky but works. (''n.b. actually this source code change probably has no effect other than to mask the warning message for v.surf.rst. It seems to be used in v.vol.rst though'')		Tricky but works. (''n.b. actually this source code change probably has no effect other than to mask the warning message for v.surf.rst. It seems to be used in v.vol.rst though'')
	: Update I have a patch which I currently test to replace MAXPOINTS with a new npmax parameter - Markus [[User:Neteler\|Neteler]] 12:56, 7 April 2009 (UTC)		: Update I have a patch which I currently test to replace MAXPOINTS with a new npmax parameter - Markus [[User:Neteler\|Neteler]] 12:56, 7 April 2009 (UTC)

⚠️HamishBowman: /* Working with contour lines */

2010-02-27T17:14:30Z

Working with contour lines

← Older revision		Revision as of 17:14, 27 February 2010
Line 69:		Line 69:
	You may prefer to try the {{cmd\|r.surf.contour}} or add-on {{AddonCmd\|r.surf.nnbathy}} modules if starting with contour line data.		You may prefer to try the {{cmd\|r.surf.contour}} or add-on {{AddonCmd\|r.surf.nnbathy}} modules if starting with contour line data.

	* See the [[Contour lines to DEM]] wiki page for a study of different interpolation methods.		* See the [[Contour lines to DEM]] wiki page for a study of different interpolation methods.

	=== Working with huge data sets ===		=== Working with huge data sets ===

⚠️HamishBowman: /* Working with contour lines */

2010-02-27T17:14:06Z

Working with contour lines

← Older revision		Revision as of 17:14, 27 February 2010
Line 68:		Line 68:

	You may prefer to try the {{cmd\|r.surf.contour}} or add-on {{AddonCmd\|r.surf.nnbathy}} modules if starting with contour line data.		You may prefer to try the {{cmd\|r.surf.contour}} or add-on {{AddonCmd\|r.surf.nnbathy}} modules if starting with contour line data.

			* See the [[Contour lines to DEM]] wiki page for a study of different interpolation methods.

	=== Working with huge data sets ===		=== Working with huge data sets ===

⚠️HamishBowman: /* Troubleshooting */ wikipedia link

2010-02-27T08:49:01Z

Troubleshooting: wikipedia link

← Older revision		Revision as of 08:49, 27 February 2010
Line 85:		Line 85:
	** Solution 2: Create the surface in multiple passes. First isolate the low density points and use {{cmd\|v.hull}} to make a polygon around them. Next run {{cmd\|v.surf.rst}} for those points to make a smooth surface between them. Then you can either run {{cmd\|v.to.rast}} on the hull area and create new random points with the {{cmd\|r.random}} module, or sample the new surface with the {{AddonCmd\|v.random.cover}} add-on script. Finally use {{cmd\|v.patch}} to create a new points file with a less dramatic change in point density.		** Solution 2: Create the surface in multiple passes. First isolate the low density points and use {{cmd\|v.hull}} to make a polygon around them. Next run {{cmd\|v.surf.rst}} for those points to make a smooth surface between them. Then you can either run {{cmd\|v.to.rast}} on the hull area and create new random points with the {{cmd\|r.random}} module, or sample the new surface with the {{AddonCmd\|v.random.cover}} add-on script. Finally use {{cmd\|v.patch}} to create a new points file with a less dramatic change in point density.

	* Spline surfaces do not handle hard breaks in the topography very well, such as found on a beach at the base of a cliff. "~~Ringing~~" may be introduced and the module may complain about "overshoots", as the surface is bent away from an input point further than the module is happy with. Users familiar with digital electronics may recall the similar problem of fitting a sinusoid to a square wave.		* Spline surfaces do not handle hard breaks in the topography very well, such as found on a beach at the base of a cliff. "{{wikipedia\|Ringing_artifacts}}" may be introduced and the module may complain about "overshoots", as the surface is bent away from an input point further than the module is happy with. Users familiar with digital electronics may recall the similar problem of fitting a sinusoid to a square wave.

	=== ToDo ===		=== ToDo ===

⚠️HamishBowman: /* Troubleshooting */ links

2010-02-26T09:55:21Z

Troubleshooting: links

← Older revision		Revision as of 09:55, 26 February 2010
Line 81:		Line 81:
	=== Troubleshooting ===		=== Troubleshooting ===

	* If your input data points include a dramatic change in point density across the region of interest you may get square artifacts in the output map. This is especially visible in a slope map or in contour lines created from the elev map with r.contour.		* If your input data points include a dramatic change in point density across the region of interest you may get square artifacts in the output map. This is especially visible in a slope map or in contour lines created from the elev map with {{cmd\|r.contour}}.
	** Solution 1: increase the ''npmin'' parameter		** Solution 1: increase the ''npmin'' parameter
	** Solution 2: Create the surface in multiple passes. First isolate the low density points and use v.hull to make a polygon around them. Next run v.surf.rst for those points to make a smooth surface between them. Then you can either run v.to.rast on the hull area and create new random points with the r.random module, or sample the new surface with the v.random.cover ~~addon~~ script. Finally use v.patch to create a new points file with a less dramatic change in point density.		** Solution 2: Create the surface in multiple passes. First isolate the low density points and use {{cmd\|v.hull}} to make a polygon around them. Next run {{cmd\|v.surf.rst}} for those points to make a smooth surface between them. Then you can either run {{cmd\|v.to.rast}} on the hull area and create new random points with the {{cmd\|r.random}} module, or sample the new surface with the {{AddonCmd\|v.random.cover}} add-on script. Finally use {{cmd\|v.patch}} to create a new points file with a less dramatic change in point density.

	* Spline surfaces do not handle hard breaks in the topography very well, such as found on a beach at the base of a cliff. "Ringing" may be introduced and the module may complain about "overshoots", as the surface is bent away from an input point further than the module is happy with. Users familiar with digital electronics may recall the similar problem of fitting a sinusoid to a square wave.		* Spline surfaces do not handle hard breaks in the topography very well, such as found on a beach at the base of a cliff. "Ringing" may be introduced and the module may complain about "overshoots", as the surface is bent away from an input point further than the module is happy with. Users familiar with digital electronics may recall the similar problem of fitting a sinusoid to a square wave.

⚠️HamishBowman: /* Working with huge data sets */ links

2010-02-26T09:53:10Z

Working with huge data sets: links

← Older revision		Revision as of 09:53, 26 February 2010
Line 73:		Line 73:
	Interpolation large DEMs from large amounts of vector points requires lot of RAM. Some hints		Interpolation large DEMs from large amounts of vector points requires lot of RAM. Some hints

	* don't create point topology (-b flag in several modules like r.random, v.in.ascii etc)		* don't create point topology (-b flag in several modules like {{cmd\|r.random}}, {{cmd\|v.in.ascii}} etc)
	* use segmentation (e.g., segmax=10 and npmin=120)		* use segmentation (e.g., segmax=10 and npmin=120)
	* interpolate in chunks with overlap		* interpolate in chunks with overlap