Localizing to Find Boundary Corners

I am going to assume that your control was used to tie in those 2 located boundary corners so now you have control coordinates on those 2 corners. Then I assume that you have record coordinates (of some sort) all around that map which includes the 2 corners that you located. If that was my job I would inverse the 2 located coords and compare the inverse between the map point coords. If the distances are close (+/- 0.5') then I would translate and rotate that map to my control using the 2 located points. Now I will have coordinates on the map record points to search for. I almost never rotate or translate my control in the field (never if I am using GPS). Later, in the office if/when I find that I have 2 solid points on a line on a recorded map, I will rotate everything to that bearing for a "Basis Of Bearings" for my map.

Assume your own control, locate the 2 boundary corners, then translate and rotate all of your projected plat corners to your control. Translating and rotating your control to the boundary corners only works if the corners are perfect. If not you have a bad rotation and a bad scale factor. 

Posted by: @field-dog

↑

My party chief does not use control points but holds two located boundary corners and localizes to them.

Edit

This is incorrect. He holds one located boundary corner. The pair he uses, calculated and located, are kept in a localization list on the data collector.

Posted by: @landbutcher464mhz

↑

I am going to assume that your control was used to tie in those 2 located boundary corners so now you have control coordinates on those 2 corners.

Unless your definition of control is different than mine, the only control we have is the point that the base station was setup on. That was the only point we set and it's a nail and disc. My definition of control is a network of adjusted points that have been established using either GNSS or total station traverse, the later tied into at least 1 NGS, GIS, or another significant monument.

Posted by: @landbutcher464mhz

↑

Then I assume that you have record coordinates (of some sort) all around that map which includes the 2 corners that you located.

Why would I locate record coordinates? The only record coordinates we used are those on the CCR #74815 (REC. 2" AXLE WITH "X" CUT S.W. COR. SEC. 12, T-23 S, R-27 E). I'm unable to attach the CCR to this post. You can look it up on LABINS (LABINS).

Posted by: @lurker

↑

Translating and rotating your control to the boundary corners only works if the corners are perfect.

But it will get us close enough to recover the boundary corners. For some reason the recovered points are about 4' south of the calculated points.

Posted by: @field-dog

↑

Unless your definition of control is different than mine, the only control we have is the point that the base station was setup on. That was the only point we set and it's a nail and disc. My definition of control is a network of adjusted points that have been established using either GNSS or total station traverse, the later tied into at least 1 NGS, GIS, or another significant monument

Yes that is what I consider control and I never make any changes to it in the field. Before I go into the field I run sets of assumed coords (kept in groups like #1-19 for MAP.ABC, #21-#49 for ROS.XYZ, etc) around any reference maps and they are on a layer called RECORD. Then when I find a couple points that are on one of those maps and tie them to my control I can translate and rotate that particular map's RECORD points to my control. Then any other points that are found on that particular map should be within 0.5' of the calc'd position or something is wrong with how I brought that particular map into my coord system or like you suspect, the point is out of position.

Posted by: @field-dog

↑

But it will get us close enough to recover the boundary corners. For some reason the recovered points are about 4' south of the calculated points.

You have no way of knowing these points are 4' south of calculated points because you have no control. You need to work on  a coordinate system that reflects ground distances as opposed to one that accounts for the scale needed to make in exact property corners match the record. An assumed system is fine. A bastardized system where you accept a found corner as being at the exact location you have calced will produce results that are not correct. So he assigned some coordinate value to the property corner you found and all the other corners were assigned coordinate values based on the record that was used to calc those positions. When the corners are not at those positions but you treat them as if they were, you end up with a screwed up coordinate system. It is no different than going back to a site where you have established control and occupying a point that is .5' away from your actual control point. All the data you collect from this setup is wrong and will need to be corrected. Treating property corners as if they have been set exactly as stated in the record will cause the same problem.

First I'll preference this by saying that I rarely use localization/transformations/calibration for boundary searches, I prefer to translate and rotate the record drawing to my system.  One simple reason being it requires separate transformations for each recorded basis of bearings.  If you have just one record drawing you're working from that might not be an issue.  Localizations were historically used to conform GNSS derived geodetic coordinates to terrestrial established local control.  They incorporate any error from both the existing local control network and the GNSS observations.  That is not a bad thing for good control but can get ugly really quick when you are holding theoretical coordinates from a record drawing.

If I were to localize onto a record drawing I would probably go with this approach.

1) Recreate the record drawing in CAD with an assumed coordinate system.

2) Recover and tie multiple monuments referenced in the drawing.  I would shoot for 4, but a minimum of 3 spread across the project are.  Since you are just looking for search coordinates I would not worry too much about boxing in the project.

3) Compare record inverses vs measured inverses to verify that you have in-fact tied the correct monuments.  I would venture to say your mysterious 4' is a result holding the wrong monuments or a bad RTK fix.

4) Perform the localization on your DC using at least one of the tied monuments out of the calculations.

5) Compare the localized coordinates to record coordinates on points you held out of the calculations.  This should give you an idea on the localization's quality.  You would definitely see your 4' hear.

6) Use the localized system to recover the needed monuments.

7) Git rid of the bastardized localization.  It will cause nothing but trouble in the future.

Like I said, I prefer to translate and rotate the record information into my control network.  I you have a pretty good feeling for the C.F. in the area you can even scale the record as needed.

If anyone is curious, here's the coordinate file. Note that the 500 number point range is the calculated points and the 200 number point range is the recovered points. We still have more points to recover.

Posted by: @john-putnam

↑

Like I said, I prefer to translate and rotate the record information into my control network.  I you have a pretty good feeling for the C.F. in the area you can even scale the record as needed

Exactly what I have been doing for years.

If you are using only two points, be sure they span the majority of the project length, so that slight errors are not leveraged into big ones.

The most important thing is to produce something that any halfway intelligent surveyor can follow for the near future with a high degree of certainty.

Speaking of those surveyors who have not been able to become halfway intelligent to date, reminds me of a fellow who called me who informed me my elevations were off by roughly four feet.  Hmmm.  Our elevations were tied to several city and DOT monuments with great agreement.  No work had been done onsite to change any of the fixed points that could be checked into.  Another fellow contacted an excellent local surveyor and informed him that his elevations were off by something like 150 feet.  Yes, idiots like that still roam the Earth.  In case anyone says maybe there was a significant datum change, that amounts to about 0.4 feet not four feet or 150 feet.

My handheld GPS / phone says so.  /s

Posted by: @field-dog

↑

If anyone is curious, here's the coordinate file. Note that the 500 number point range is the calculated points and the 200 number point range is the recovered points. We still have more points to recover.

I assumed that the data you call recovered is what you measured with your GPS on your control network. If that assumption is correct then your survey is good. All the found (recovered) points (#201 to #217) fit the adjusted PLAT coordinates (within 0.2') except for 2 points. #207 is W'ly 1.5' but your notes say its leaning and #215 is S'ly 0.9' but not the correct description.

Here is my adjustment. I held the 4x4 mon's at #216 and #217. I also made a duplicate point at #501 numbered #522 to be included in the adjustment so I could keep #501 out and unadjusted. I translated & rotated all the PLAT calc'd points (#502 thru #552). Translated from #537 to #216 and rotated at #537 using old bearing = (#537 to #529) to new nearing = (#216 to #217).
Everything is now on your control network coordinates based on #201 to #217. The txt file is tab delimited like your was.

Posted by: @landbutcher464mhz

↑

Everything is now on your control network coordinates based on #201 to #217

I cannot figure out how to edit my post so I'll use this new post. That sentence I quoted above should end with based on #216 to #217.

Sorry, maybe someone can tell me how to edit one of my posts. It appears that there is a time limit?

Occupy existing GPS control with a projection imbedded in your DC, Put everything on it.

You can occupy the CCR monument assume it's good, then check it's real-world location and adjust if needed. I wouldn't trust it until I know it's good. I do that with my own control. That doesn't have to impede your initial field work, the adjustments can be done in the office later. 

Not sure what you mean by localization. I only project coordinates with a defined system. The basis of coordinates should be clean, no adjustments to it once defined.

Coordinate systems are always moving, small slips happen over time so no control point will be fixed in the global system for long. It's best to know the time attached to the coordinate. 

It pays off so much to keep everything on the same system for an area, Still, there will be differences between old and new control. We are seeing about 0.20' on recent projects from our mid 1990's to mid 2000's era control and the same points occupied the last two years. No doubt in California or Hawaii those shifts are much larger.

@john-putnam said it best. You can get into deep trouble with what is being done. It is the equivalent of taking any plat and computing coordinates on it and then randomly selecting 2 points and doing a resection off of those computed points without knowing the measured distance between any of those monuments. Translate rotate the look clacked points to your system or geo reference the map or translate that to your system to find the other corners. Once he localized or calibrated that base station coordinate is changed it is no longer what it was before the localization/ calibration. And technically if he only used 2 he is just translating and rotating. So take those same computed points and just translate rotate them to your measured points not the other way around. Be very careful as I have seen this done before and it bit some good surveyors in the rump. 

Others have touched on the mathematical reasons for translating and rotating, I'll add that there's value in visualizing the discrepancy between field measured control or boundary markers and record control or boundary. Any time you're tying to control, regardless of whether it's a benchmark or an boundary corner, you want to see the differences in angle and distance. Even if you have a strong background in statistics, it can be difficult to isolate the error in a localization.  So, it's a good thing to translate and rotate and see that you're missing by four feet because it signals that there may be a problem that needs to be addressed before it turns into a calamity.

Even if your primary task is to topo a sidewalk, finding a control point or even a boundary monument that is incorrectly represented in plans or the record books can save your PLS lots of time and money. Live up to the maxim, "I'm a surveyor, I find all the problems you didn't know you had".