[Proj] Proj4js EPSG:4326 to EPSG:26786 (SP NAD27) 100+ feet Off

Scott DeHainaut scottdehainaut at comcast.net
Sun Oct 16 13:37:17 EST 2011


Mikael,

The Proj4js definition files are a bit of a black box for me but I used your
latest transformation and was able to reproduce your 0.53 error which tells
me I'm on the same page.  The area for which this application is focused is
fairly specific within the State of Massachusetts, about 144 square miles
(approx 12x12 miles).  Another (admittedly low-tech) approach I was
considering was 1) round the coordinate readout for the user to the nearest
foot and 2) simply introduce an appropriate X and Y offset to the original
transformed coordinates.  I've used several other known survey points and
determined an average horizontal offset of -134.98 and vertical offset of
-815.74.  When I add these offset values to three different sets of
transformed coordinates my average distance error is 0.32 from their
surveyed locations.  As I sample additional known points across the
application's region I'll be curious to see if this error remains
consistent.  Both this method and your "+towgs84" method are producing
distance errors under 1 foot which, for this application, are acceptable
given the final component of my approach is to 3) include a standard
disclaimer that the results are not intended to replace engineering-class
field survey methods.

Many thanks for your help with this!

-Scott DeHainaut


Mikael Rittri wrote:
> 
> PS:
> Here is another transformation from NAD27 to NAD83(86), designed for
> Massachusetts:
> 
>    +towgs84=-38.67,117.84,209.59,0.0091,-0.9236,2.5411,-1.75953
> 
> which I think is better, but use it at your own risk. 
> It is expressed as a Position Vector Transformation, as Proj.4 wants it. 
> 
> I got it from the SEVENPAR program, by entering six sample points that I
> had transformed with an online NADCON calculator. 
> The max horizontal error for my six sample points was 37 cm (1.21 ft).
> For your test point, the horizontal error is 16 cm (0.53 ft). I don't know
> the maximal error in the whole state. 
> 


--
View this message in context: http://osgeo-org.1803224.n2.nabble.com/Proj4js-EPSG-4326-to-EPSG-26786-SP-NAD27-100-feet-Off-tp6894605p6898300.html
Sent from the PROJ.4 mailing list archive at Nabble.com.


More information about the Proj mailing list