I'd chain/traverse at least one of those lines (unless it's hillacious). Then you'd know. For 10 ft you don't have split the atom.
FrozenNorth, post: 454796, member: 10219 wrote: If you take a look at your measurements in SPC with no calibration applied, you could at least see whether your calibration is completely fubar and eliminate that as a factor. Even with the distortion in your SPC zone, you're only going to be seeing a foot or less per mile, not 10 feet.
Exactly, that why I'm thinking it is a CAD error.
THiggins, post: 454794, member: 7840 wrote: A two point calibration will hold two points and scale all measured distances between them, as well as rotating to those two points. If you specify within your software that you want to hold a scale factor of 1.00000, then it will not scale, and you will see residuals on your localization.
If you have a scale factor applied, you still will not see residuals with only two localization points as there aren't any; your localization has held the record locations as gospel and forced your observed points to match with no residual.
As to the larger question; what do other record maps say? The latest was 1996, but surely you have older maps in the file that would prove out a distance one way or another, or confirm that you're observing the correct monument.
The older records are 1875 GLO records and the distances vary over 40 feet. Pretty hard to correlate anything using them.
There were no residuals shown, The scale factor shown is 1.00361357. Max H. Residual 0.000sft. That's 9.5 feet. I believe that's where the problem is.
Skeeter1996, post: 454807, member: 9224 wrote: The older records are 1875 GLO records and the distances vary over 40 feet. Pretty hard to correlate anything using them.
There were no residuals shown, The scale factor shown is 1.00361357. Max H. Residual 0.000sft. That's 9.5 feet. I believe that's where the problem is.
When I hold the scale factor to one, I am getting a residual on one point of 4.848 feet. I think you're on to something. Thanks for the tip THiggins
Wait, are you saying that you're calibrating/localizing to two points from a 20-year old survey plat? I thought maybe you meant local control points. How would you ever make independent measurements of a line if you just calibrate/localize to the old plat?
FrozenNorth, post: 454817, member: 10219 wrote: Wait, are you saying that you're calibrating/localizing to two points from a 20-year old survey plat? I thought maybe you meant local control points. How would you ever make independent measurements of a line if you just calibrate/localize to the old plat?
Nope I'm localizing to a 1996 survey. The GLO measurements don't come close to what conditions are at the site today. The existing monuments are placed at road intersections and pipe monuments of unknown origin. The 1996 survey tied them and a record was produced.
Why the hell are you "calibrating/localizing" in the first place?
This is Survey 101!
Loyal
I never liked localization to any survey including my own until I was sure that the corners had not been moved. A 2 point localization would translate, rotate and scale to the 2 points used. If that distance is incorrect, then localization matches to the incorrect distance. You should be holding one point and just rotating to the other. If the points match pretty close then you can hold them. I would also use more that 2 points the same could happen to the bearing holding only 2 points.
Skeeter1996, post: 454819, member: 9224 wrote: Nope I'm localizing to a 1996 survey. The GLO measurements don't come close to what conditions are at the site today. The existing monuments are placed at road intersections and pipe monuments of unknown origin. The 1996 survey tied them and a record was produced.
Why would you force your measurements onto a 20 year old survey that is not your own? Does this methodology pass for boundary resolution in your area?
Skeeter1996, post: 454819, member: 9224 wrote: Nope I'm localizing to a 1996 survey. The GLO measurements don't come close to what conditions are at the site today. The existing monuments are placed at road intersections and pipe monuments of unknown origin. The 1996 survey tied them and a record was produced.
Localizing is a garbage-in, garbage-out operation. It's one thing to localize to site control that's already known to be nuts on. But what you're doing is bad juju my friend. Let the fellas here set your path straight.
Loyal, post: 454822, member: 228 wrote: Why the hell are you "calibrating/localizing" in the first place?
This is Survey 101!
Loyal
At the risk of revealing trade secrets. I precalculate corner positions for every corner I need to tie into to complete my survey. I overlay those points over an aerial photo as close as I can using a USGS quad section corner to get me close. The quad and aerial are georeferenced so they are close thought not exact. Then I load all those points into my GPS Data Collector that I have calculated and label them search. Then I head out to the survey project. If I'm lucky enough to find one corner I GPS it and then do a one point site calibration with the calculated position to the found position. That translates my calculated positions into the ground positions. I start searching for corners using my calculated positions as search points using the Layout Function. The second point is the tricky one to find since there is a rotation element still in my calibration that is different from the field. Once I find a second point I tie it and add it to the site calibration. This rotates me into the basis of bearing system of the old surveys. I have to remember to set the scale factor to 1.00 or the site calibration will apply a scale factor to all my calculated search points. Then I'm off and running using the Layout function in the data collector. I usually find the corner I'm searching for within a reasonable metal detector search area around my staked out search area. If it's a primary control point I add it to my site calibration. If it's within a couple of tenths, I apply that calibration. The more good calibrations I can apply the more refined my search points become. Usually an old Theodolite EDM survey will be off around 5 feet or so because they didn't least square adjust their survey. I've found GPS surveyed points in projects that are within a half of foot of their calculated position and then all of a sudden one will be 9 foot out of position of it's calculated position. 9 feet seems to be a number that shows up quite frequently.
That's why I do site localizations/calibrations. The aerial photo backdrop on my search map shows me routes to get around the project. If it's a multi day project I adjust the aerial photo by using road intersections and fence corners to align it better to what's on the ground. The aerial photo is also invaluable to explain the survey to the client. They can relate much better to a picture of their land than a drawing with a bunch of lines.
billvhill, post: 454827, member: 8398 wrote: I never liked localization to any survey including my own until I was sure that the corners had not been moved. A 2 point localization would translate, rotate and scale to the 2 points used. If that distance is incorrect, then localization matches to the incorrect distance. You should be holding one point and just rotating to the other. If the points match pretty close then you can hold them. I would also use more that 2 points the same could happen to the bearing holding only 2 points.
If you don't hold your scale factor to 1.00 the 2 point calibration will scale your distances and will not report the residuals for you to evaluate. Site calibration can be a dynamic process. You can be constantly adding and deleting calibration point to improve the residuals. I have never like 4 point calibration because nobody has ever been able to explain what the calibration routine is doing. I heard one explanation that it rubber sheets things. What the heck is that?
FrozenNorth, post: 454843, member: 10219 wrote: Localizing is a garbage-in, garbage-out operation. It's one thing to localize to site control that's already known to be nuts on. But what you're doing is bad juju my friend. Let the fellas here set your path straight.
Sometimes old surveys are garbage. Site calibration allows you evaluate them and determine what quality they are. If you only do 2 point calibrations holding the scale factor to 1.00 they will only translate and rotate your survey to match the old survey. Most old surveys have a pocket compass basis of bearing. Your GPS bearing can be as much as 26 degrees different. Makes it pretty difficult to search for old pins.
roger_LS, post: 454837, member: 11550 wrote: Why would you force your measurements onto a 20 year old survey that is not your own? Does this methodology pass for boundary resolution in your area?
Well the original monuments on the ground determine the boundaries, not a least squared adjusted OPUS corrected GNSS survey. Finding those old corner positions is the most important part of boundary resolution.
THiggins, post: 454794, member: 7840 wrote: A two point calibration will hold two points and scale all measured distances between them, as well as rotating to those two points. If you specify within your software that you want to hold a scale factor of 1.00000, then it will not scale, and you will see residuals on your localization.
If you have a scale factor applied, you still will not see residuals with only two localization points as there aren't any; your localization has held the record locations as gospel and forced your observed points to match with no residual.
As to the larger question; what do other record maps say? The latest was 1996, but surely you have older maps in the file that would prove out a distance one way or another, or confirm that you're observing the correct monument.
I forgot to hold the scale factor to 1.00 in my 2 point calibration. That is the problem. Now I'm trying to defend why I even do site calibrations. Thanks for the tip!
Seems like you have a reasonable search procedure, but a lot of readers initially thought you were talking about real measurements, which need to be done without being distorted by calibrations.
As mentioned above, though, if one of the record measurements is significantly off the calibration can send you searching a longer way off.
Skeeter1996, post: 454852, member: 9224 wrote: At the risk of revealing trade secrets. I precalculate corner positions for every corner I need to tie into to complete my survey. I overlay those points over an aerial photo as close as I can using a USGS quad section corner to get me close. The quad and aerial are georeferenced so they are close thought not exact. Then I load all those points into my GPS Data Collector that I have calculated and label them search. Then I head out to the survey project. If I'm lucky enough to find one corner I GPS it and then do a one point site calibration with the calculated position to the found position. That translates my calculated positions into the ground positions. I start searching for corners using my calculated positions as search points using the Layout Function. The second point is the tricky one to find since there is a rotation element still in my calibration that is different from the field. Once I find a second point I tie it and add it to the site calibration. This rotates me into the basis of bearing system of the old surveys. I have to remember to set the scale factor to 1.00 or the site calibration will apply a scale factor to all my calculated search points. Then I'm off and running using the Layout function in the data collector. I usually find the corner I'm searching for within a reasonable metal detector search area around my staked out search area. If it's a primary control point I add it to my site calibration. If it's within a couple of tenths, I apply that calibration. The more good calibrations I can apply the more refined my search points become. Usually an old Theodolite EDM survey will be off around 5 feet or so because they didn't least square adjust their survey. I've found GPS surveyed points in projects that are within a half of foot of their calculated position and then all of a sudden one will be 9 foot out of position of it's calculated position. 9 feet seems to be a number that shows up quite frequently.
That's why I do site localizations/calibrations. The aerial photo backdrop on my search map shows me routes to get around the project. If it's a multi day project I adjust the aerial photo by using road intersections and fence corners to align it better to what's on the ground. The aerial photo is also invaluable to explain the survey to the client. They can relate much better to a picture of their land than a drawing with a bunch of lines.
What trade secrets? Every survey crew I have talked to can do this, the Lecia salesman even shows you how to do it.
This is how I do my look points, I plot the survey and the adjoiners, I overlay them on an aerial, I guess where I think the property lines are, going by ROW and fences and anything else on the aerial. Go to the location I think will be the best at finding the first corner. Then I just inverse to see how off my look points are. I remember the distance and angle and just in my head move that much for the other look points.
Calibration was developed to allow the early users of RTK a way of "getting on" older, usually instrument control. It was mainly needed to help with elevations, now Geoid models are far more accurate than any calibration can be.
It became very apparent that calibrations would warp the GPS measurements and wasn't the way to go.
It's become a last resort function only to be used in the most dire circumstance.
I see what you are doing, but I would suggest there are much better ways of doing it that don't involve messing with the GPS data, warping it if you will, a two point calibration can do many things almost all of them bad.
I know the salesman sell that function, I think it's been maybe five years or so since the last time I was forced in to applying one.
GPS is a powerful tool, you can put good functional coordinates on the points located using it. These can be projected using an almost infinite selection of choices, all reproducible by anyone following your footsteps. But no one will be able to reproduce a calibration without your electronic file.
Does it really matter? Maybe you will just return the record measurements after finding the monuments.
But for me I want any survey to be on a projection, tied to NGS system with good underlying Lat., Longs., if for no one else then for my use next time I'm in the area (and it's so easy to do these days).
I prefer to calculate plats needed for my survey. I reserve a group of points for entered points, plats and calculations usually 1-99, or 1-199 and then begin my survey at 100 or 200. It is just as easy to translate and rotate each group of points then to be changing my GPS points to match the plat. I don't ever have to question whether my coordinates are the problem.
Skeeter1996, post: 454856, member: 9224 wrote: Well the original monuments on the ground determine the boundaries, not a least squared adjusted OPUS corrected GNSS survey. Finding those old corner positions is the most important part of boundary resolution.
If that's the case, you might as well just use a consumer grade handheld GPS, this would probably get you close enough to find and mark the monuments. Why the fancy equipment?