Emlid

Feature request: logic on time mark pin


(WSURVEY) #64

just checking… you know u can change the number of decimal output for .pos file times in RTKPOST [options] > [output] page? no need to go into the rinex or .ubx file to get all 7 decimals.

going over the error estimates again, i assign a standard deviation for easting as the square root, of the sum of the squares of the following:

  • 0.01m as a base uncertainty
  • timing uncertainty * E component of horizontal velocity
  • stability number (standard deviation of 4 surrounding dE)
  • Q penalty (assigned if sum of 3 surrounding Q’s is > 3. It is arbitrary 0.5m * sqrt(sum(3 Q’s))

northing stdev is assigned similarly. height same except 0.015m base uncertainty and arbitrary 0.8m for Q penalty factor. obviously with my 4ms timing, the timing uncertainty dominates the horizontal stdev in straight flight, with the stability number blowing it out in turns and in wind. I don’t care for pos file error estimates if phase fix hasn’t been held for a few epochs and that is the purpose of my Q penalty. I just tried to think of the relevant error sources and apply them in the quantities I think were appropriate. I haven’t done a post-adjustment, group variance factor test to see if the stdev’s are statistically ‘correct’, but the errors in photoscan post adjustment looks fine to my eye. and my calibration/check points hit pretty well.


#65

ah, ok, thanks!

ok, thank you, I will include that!


Does any one do lever arm correction?
(WSURVEY) #66

well, we are either both doing something right, or wrong! just checked -200ms offset on one of my flights. we agree to within ± 3mm easting and ±2mm northing and ±0.2mm height for interpolated events in WGS84 LLh.

your program did not like me not adding an extension to the output file. it just dumped the file path back in there, except with ‘txt’ between every character.


(WSURVEY) #67

Also changing the output to gps week & seconds may be easier to work with (parse and offset) than hh:mm:ss


#68

I cannot replicate that error. Did you use the choose file dialog or did you copy and paste/edit within the test field? Which extension did you choose? Could you please try to leave it to the default value and test whether it uses parts of the pos file name to generate the output file name?


(WSURVEY) #69

yes i used the choose file dialog, navigated to my preferred directory, and typed my own filename, without extension. that’s when i get long, silly string with ‘txt’ all through the name.

when left blank, the output files are generated correctly with parts of pos filename.


#70

‘.txtJ.txt:.txt/.txtN.txte.txtu.txte.txt .txtS.txtt.txtr.txtu.txtk.txtt.txtu.txtr.txt/.txtU.txtA.txtV.txt/.txtS.txtt.txta.txtn.txtd.txto.txtr.txtt.txte.txt/.txtF.txtö.txtr.txtd.txte.txtr.txtv.txte.txtr.txte.txti.txtn.txt .txtP.txte.txte.txtn.txte.txtt.txta.txtl.txt/.txt2.txt0.txt1.txt8.txt/.txtG.txtP.txtS.txt/.txtU.txtA.txtV.txt/.txtr.txta.txtw.txt_.txt2.txt0.txt1.txt8.txt1.txt1.txt1.txt0.txt1.txt1.txt2.txt6.txt_.txt0.txt0.txt2.txt_.txte.txtv.txte.txtn.txtt.txts.txt’

ok, I managed to replicate it…

I have fixed it. I do not have any error management up to now, so there are more such strange things if the input is other than expected.


(WSURVEY) #71

also, little green progress stripe continues after processing has finished.

nice work on the program by the way. it annoys me a bit that i have not learned to script to be able to produce useful standalone exe programs.


#72

Thanks! If you are resilient against all the errors and satisfied with a overcasual programming style, its astonishingly easy with Python… .

Is the progress bar not stopping even when the program finishes correctly? Same: I cannot replicate… .


#73

I have changed the settings, no I get

2018/11/10 11:27:10.2004089 in the pos file and
2018/11/10 11:27:10.1900000 in the obs file

Does RTKlib calculate a more accurate time or is there an error?


(WSURVEY) #74

i’ve noticed that too. the offset on my latest processing was about 0.00695 seconds. hopefully emlid can come to this one.


(WSURVEY) #75

yes. my last message is 'writing output file". no more confirmation that the process is finished. green progress indicator keeps moving left and right.


#76

ok, I look into that, there should be a message that the processing is finished


#77

I could imagine that the time difference is a result of a more precise calculation of the time through RTKlib? Probably the time in the observation file is based on one satellite and during the calculation by RTKlib the “real” time is derived as a the 4th unknown?


(WSURVEY) #78

it’s possible i supppose. but our GS16 Leica antenna/receiver has no problem recording epochs showing lots of zeros. and doesn’t get changed in the pos file.


#79

@emlid could you please help? Or someone else @tatiana.andreeva, @rtklibexplorer?


(WSURVEY) #80

Hi Tobias,

I’m thinking I might be able to integrate your program into my workflow. Could I get a custom version to output an extra tab delimited photoscan/metashape ‘PS_Events…txt’ file with PT, La, Lo, h, sdE, sdN, sdH? If you get the sdE, sdN and sdH fields working of course. It would need an extra input for timing accuracy (ms) on your program interface and maybe the Q penalty factor (m) if you were thinking of making that part of the error estimation.

PT would be just filled with DJI_xxxx.JPG if you didn’t specify first image?

Cheers.


(Igor Vereninov) #81

@tobias-dahms Our interpolation code is available on GitHub repo, so you can compare it with the script that you are developing.


#82

@igor.vereninov thank you very much for the hint! Could you point me to the file/line, I’m somehow lost in the repo.

I think I would have also to start to look in to c programming and I would be a great help I the one who did write the code just tells mr that I’m worng or whatever.


#83

Hi, great, I will look into that/implement the error estimation as described above at the weekend… .