Coordinate reversion?

[Ned Ferguson]

I am having a serious problem with FieldGenius that I will try to communicate as best I can. I am still trying to wrap my brain around it myself. I have a job (topographic) which I did alone using a combination of the new Topcon Robotic GPT 9005A and Topcon HiperLite RTK. Using AllegroCE with v. 3.0.1 (2006-11-27)

Day 1:
Shot points 1-244 with Robotic - no problems.

Day 2 (before lunch):
After analyzing the data in the office, I uploaded an ascii file with rotated coordinates, then imported the ascii file into FG, telling the program to overwrite the points. No apparent problems so far. I then set up the RTK and shot 4 control points in common and adjusted the RTK points to the ground coordinates. I spend the next several hours staking out calculated points looking for more evidence, some of which I found. I did some limited as-built and broke for lunch (points 259-283). Still no problem so far.

Day 2 (second half):
I set up the RTK base on the same station. I gave FG a new antenna height and continued doing the job as I had done before. The first point after lunch with the new seup is 284. At point 322, I noticed the points were not coming out on the screen where they should. Something is wrong, I muttered to myself. I staked out one of my control points. FG is telling me the point is way off from where it should be. I start cursing about this time. I look in the transformation settings at the parameters. They look the same as far as I can remember (the SF had not reverted to 1.000 as I hade expected) but the residuals are now way off. This happened by itself. I figured FG has somehow reset the transformation parameters, but I really did not know how.

I attempted to fix the problem by simply clicking "calculate parameters" again. The risiduals now look good again. Figuring this worked OK, I staked out another control point just to be sure. Sure enough, the control point stakes out OK. Now satisfied, I continue with the rest of the job thnking everything is a-ok.

Back at the office:
I upload points to my computer, thinking that I only have a small group (284-322) that will be out of place. On the contrary, I seem to have points all over the place. In fact most of the job is rotated incorrectly. This is inexplicable to me. Knowing that I had staked caclulated points in the field, and that they came out exactly where I had expected, I could not understand how this could possibly be.

I examined the raw data. All the SP point records show up correctly, verifiing that I uploaded the rotated points correctly.

After looking carefully at the raw data and comparing the coordinates now in the data collector, it seems that I was incorrect in assuming the RTK transformation parameters were off.

Now read carefully.

What had actually happened is that the coordinates that I had uploaded earlier had reverted back to their old values by themselves. By re-transforming the RTK points, I actually put the job back on the assumed rotation I had sarted with on the first day. This apparently happened somehow when I began the new setup after lunch.

I now have several groups of points with mystery rotation that I am trying to figure out what to do with. There is no evidence of any of this "reversion" occuring in the raw data. I do not want to have to re-shoot the job. This little quirk has cost me untold hours of grief. I am still trying to figure out what happened. I just hope I can.

[Ned Ferguson]

The good news is that writing all that down helped me to formulate a plan and realize what happened. I recovered the data. This seriously took the wind out of my sails today.

The "reversion" theory is still the only one that makes sense to me. The pre-lunch coordinates and observations were all correct. The post-lunch coordinates were wrong.

[Jason Poitras]

Hi Ned,

Could you send me your FieldGenius project so I can examine it.

I have two questions:

Did you do any raw file editing in FieldGenius and then recoordinated it?

Did you use the transform functions which is accesible from the coordinate editor?

Another suggestion I have so you get faster service is to go to www.microsurvey.com/helpdesk and create a support ticket. Posting technical support questions on our user forum isn't always the best place to get immediate answers to your questions.

[Ned Ferguson]