[GRASS-dev] i.points, i.vpoints now use G_parser()

[Michael Barton]

So I should display a forward RMS and a backward RMS.

For each point, should I just display a single error distance or try to
display reverse for raster and forward for vector?

Michael
__________________________________________
Michael Barton, Professor of Anthropology
School of Human Evolution & Social Change
Center for Social Dynamics & Complexity
Arizona State University

phone: 480-965-6213
fax: 480-965-7671
www: http://www.public.asu.edu/~cmbarton



> From: Glynn Clements <glynn at gclements.plus.com>
> Date: Tue, 27 Jun 2006 16:25:43 +0100
> To: Michael Barton <michael.barton at asu.edu>
> Cc: Hamish <hamish_nospam at yahoo.com>, <grass-dev at grass.itc.it>
> Subject: Re: [GRASS-dev] i.points, i.vpoints now use G_parser()
> 
> 
> Michael Barton wrote:
> 
>> For total RMS, do I simply combine forward and backward (i.e., like I have
>> twice as many points)? Or is it better to display separate forward and
>> backward RMS values? Same with displaying 'g' for individual points--just
>> display forward or display the average of forward and backward?
> 
> Forward and backward errors will be in different coordinate systems,
> so it isn't meaningful to compute an average (arithmetic mean).
> 
> For a projection which consists solely of translation, rotation and
> uniform scaling, the forward and backward errors will be directly
> proportional.
> 
> For projections with significant levels of distortion, the forward and
> backward errors will differ depending upon the scale factor in the
> direction of the displacement.
> 
> E.g. if you had an image of the whole earth, misplacing a GCP near the
> horizon (where the surface is tangential to the view) would result in
> a large forward error but a small backward error, while misplacing a
> GCP in the centre (where the surface is perpendicular to the view)
> would result in the forward and backward errors being roughly
> proportional.
> 
> Also, bear in mind that rasters are projected using the reverse
> transformation, while vectors use the forward transformation.
> 
> -- 
> Glynn Clements <glynn at gclements.plus.com>