> > > New Mexico East of the Rio Grande
> various zones of the Texas Coordinate System extend perfectly into the part of New Mexico
Perfectly?
I guess you could extend your Lambert zones 4201 & 4202 West to the Rio Grande. You would only be working in negative Eastings for the last 140 miles.
> > various zones of the Texas Coordinate System extend perfectly into the part of New Mexico
>
> Perfectly?
>
> I guess you could extend your Lambert zones 4201 & 4202 West to the Rio Grande. You would only be working in negative Eastings for the last 140 miles.
Well, negative Eastings shouldn't bother any of the software used in Texas. It's a small price to pay for helping y'all back into the Republic of Texas. :>
Localizations include two possible variations (or a combination of both).
One part is to determine a projection. Those that I have been able to find documentation on thus far (Trimble and Javad) use Oblique Stereographic projections. This is necessary if the user is beginning from a geographic coordinate rather than a grid coordinate in an already defined projection (such as State Plane), as the Helmert transformation (second part) requires Cartesian coordinates. Because this process involves matching a geographic coordinate to a Cartesian coordinate (lat/long = n/e) it is important to maintain the same reference for the source of geodetic coordinates. If a localization is built from a "here" (autonomous) base position, as is often the case, the localization is tied to that geographic coordinate or surveyed coordinates based on that geographic coordinate forever (or until the localization is recomputed). Introducing another "here" position or introducing any geographic coordinate will cause huge problems on the order of the error between the geographic coordinates (i.e. several meters). And as mentioned, what is the projection oriented to? The central meridian of the projection, with convergence applicable to coordinates East or West of the meridian.
The second part is a Helmert 7 parameter transformation between target values (imported from an already established survey) and surveyed values (expressed in the new projection or in the already defined projection). The results of the transformation are the localization parameters which include Rotation in three axes (around the U axis for North orientation and around the N axis and E axis for tilt); Translation in three dimensions (NEU); and scale (a single scale factor).
In my opinion, the second part is where users drop the ball and get burned. Helmert finds the best possible parameters to provide the lowest possible difference in target and surveyed coordinates (residuals). Helmert can cover a multitude of sins, particularly with only a few points. Two points will always provide 0 residuals. This is because any error between target and surveyed will be scaled out, even the gross errors. So the user needs to look carefully at his scale factor and have a good idea of what he should be expecting. If the surveyed system is State Plane and the target system was measured along the ground, then he should expect a scale factor near the combined factor for the points in the localization. If the surveyed system and target system are referenced to the same coordinate system (such as State Plane) the scale factor should be at or near perfect unity. Tilt (or inclination) can also be a mess. If the geometry isn't right for the intended use of the localization, then significant error can be introduced. Consider a three points (minimally necessary to create a plane) in which two points are 1000 feet apart and the third lies halfway between and only 10 feet from the line between the extremes (a very skinny isosceles triangle). If the middle point is off in elevation by a 0.10 of a foot (at a 10 foot offset) how far off will this plane be when projected 100 feet from the baseline? 1 foot! Use of inclination/tilt should be done with great care. The rotation from North should also be considered carefully. The orientation of the underlying survey projection will be geodetic at the central meridian, whether the projection is new, as from Step 1 above) or existing, as in State Plane. If localizing to a target system with a different bearing relation (such as a compass oriented total station survey) the localization is no longer going to be related to Geodetic North at the central meridian. It is now related to the target system. If the survey system and target system are the same, then the rotation solved by the localization should be very near zero.
If the target coordinate system and surveyed coordinate system are identical, I would highly recommend (subject to a very well reasoned objection) that rotation, scale and inclination be forced to 0"/0ppm. It makes no sense to redefine the foot or redefine North.
Localizations do a lot of hard work fast and can be very useful for quickly verifying the relationship between surveyed points and established points. Documentation of localizations, knowing what to expect before pressing the solve button. Localizations will generally result in two distinct outputs: a projection (existing or new) and a set of transformation parameters. Both should be documented, and both should be carefully evaluated by a knowledgeable professional before being used for production.
I tried for a few years to use calibrations, they were always quite clearly inferior to coordinate projections.
I finally decided to rotate the existing coordinates to my superior GPS system when I could and do calibrations only as a last resort.
As far as using some difficult to define projection around some here point that really makes me scratch my head.
Why is that an acceptable procedure? I really don't understand the reasoning behind doing it. It's so simple to define a projection for the job, you don't have to be in the field to do it, heck you know where you are going (I hope) before you get there.
I have run into people that go out every morning on the job site and calibrate. It of course introduces a new calibration each day.........
