[Qgis-user] Trimble GeoXT 2005 Accuracy
kirk
kirk at nortekresources.com
Tue Mar 9 04:32:25 PST 2021
Hi SpringerI just looked up the marine beacon specs and they are designed for 10m accuracy. This may explain why your non dgps data is closer to your survey corner.Sent from my Galaxy
-------- Original message --------From: Springfield Harrison <stellargps at gmail.com> Date: 2021-03-08 3:17 p.m. (GMT-04:00) To: Nicolas Cadieux <njacadieux.gitlab at gmail.com> Cc: kirk <kirk at nortekresources.com>, Jorge Gustavo Rocha <jgr at geomaster.pt>, qgis-user at lists.osgeo.org, Greg Troxel <gdt at lexort.com>, Dan <19dmb83 at gmail.com> Subject: Re: [Qgis-user] Trimble GeoXT 2005 Accuracy
I'm resending this without the map as there is a size limit. The moderator may let it through, I hope . . . .
Hi Nicolas, thanks for your observations. I'll try to answer your questions, please see the attached map, especially Map A:
Note that my previous email contained information for Map B; Map A is based on the Municipal Cadastre (NAD83 UTM zone 10N) and illustrates the problem as well. Other locations based on Provincial Monuments and/or the Municipal cadastre (not illustrated here) have yielded similar results.
How many “known” points have you tested?
2 in this case, Maps A and B
Also several other locations with similar results
How where those point position calculated.
From the Municipal cadastre, visible in Map A
Make sure the coordinates are in the right CRS
NAD 83 UTM 10N used throught. See workflow in previous email
When converting from the monument’s CRS to NAD83 UTM zone 10N, are you using the correct grid files?
[No monuments in this example] These were brought into QGIS from the Municipal GCM database CSV (NAD83(CSRS) 3.0.0.BC.1.CRD) and reprojected by QGIS to EPSG:26910 - NAD83 / UTM zone 10N
Presumably QGIS would choose the correct grid files
Municipal Cadastre is NAD83 UTM zone 10Np, li { white-space: pre-wrap; }
Find a geodesic point that is in the middle of a field or on the side of a highway with no obstacles.
Map A is open sky
Make sur your observations will be done when the constellation is well distributed in the sky
As you probably know, TerraSync provides for PDOP, HDOP, SNR and Horizon masks to preclude collecting poor quality positions. These were set towards the "Precision" end of the scale
What post processing techniques are you using? How far is the base station from your unit?
Real time was SBAS or RTCM; Post processing using the Pathfinder Office differential correction engine, baseline about 30 km
How long are the observations? Have you tried other methods of post processing like PPP?
Logging interval is 5 sec; 33 to 2037 positions per point
Did not use PPP. This is a test of mapping best practices, not geodesy
Have you contacted Trimble?
Yes, no response
Have you looked on there site to see if there is a software update (firmware) for the unit or the post processing software?
Yes, receiver firmware is the latest, PFO and Terrasync are older but compatible
Thanks Nicolas. If I have missed something, I hope someone can point it out, I've tried to cover all the bases based on my training and experience.
-----
Cheers, Spring
On 08/Mar/2021 06:40, Nicolas Cadieux
wrote:
Hi Harrison,
How many “known” points have you tested? How where those
point position calculated. They could be off. If you are using
state geodesic monuments, try to find the documented precision
of the monument. States have different types of monuments, some
are very old and have different standards. Make sure the
geodesic point is not the problem. Make sure the coordinates
are in the right CRS. As an example, if the coordinates are
published in NAD83 original but you are assuming NAD83(CSRS),
then you have a problem. When converting from the monument’s CRS
to NAD83 UTM zone 10N, are you using the correct grid files?
What is the published precision for this reprojection?
You say you have houses and trees. This could be the
problem. Find a geodesic point that is in the middle of a field
or on the side of a highway with no obstacles. Make sur your
observations will be done when the constellation is well
distributed in the sky. I believe Trimble has a observation
planing software that can help you figure out the best time for
observation. This could explain why the GEoTX are to the east
unless the observations where made at the same time and same
conditions (ex leaf off).
What post processing techniques are you using? How far is the
base station from your unit? If you are using a state correction
service, can you select more stations? How long are the
observations? Have you tried other methods of post processing
like PPP?
Have you contacted Trimble? Have you looked on there site to
see if there is a software update (firmware) for the unit or the
post processing software?
Nicolas Cadieux
https://gitlab.com/njacadieux
Le 8 mars 2021 à 05:24, Springfield
Harrison <stellargps at gmail.com>
a écrit :
Hi Kirk,
Thanks again for the ideas.
Re "I assume your raw data files are being converted to gpx on a
computer since the raw terrasync files are proprietary binary
files". Not sure why you would make this assumption - PFO does
not export GPX files, only GIS files of many kinds, although one
could create a custom format I suppose.
I have always avoided non-GIS formats (Garmin, GPX, GDB, KML,
KMZ, GoogleEarth, iPad/Tablet "mapping", etc.). I'm in the
process of re-mapping a tablet based tree inventory using SW Maps
with a Total Station survey as many of the trees are near the
property boundary. Some of the tablet errors are quite large.
Due to the tree canopy, GPS quality is variable.
I know that many people use tablets and hiking GPS as mapping
tools but I have little faith in them for that purpose.
For many years my work flow has been: Trimble Receiver +
RTCM/SBAS -> Pathfinder Office [+ RINEX Post Processing] ->
SHP files -> GIS (QGIS or Manifold GIS). The CRS is NAD83 UTM
10N throughout, for my home area at least.
None of these steps offer any option to choose or modify the Base
Station CRS so I don't think that would be the culprit in my NW
data offset, although maybe I've missed something.
Last fall I collected quite a few points in an attempt to
quantify the problem, if that's what it is. Here are some
summaries:
Average distance from "Known" point
(m)
Location
Receiver
Correction
Corner
IP NW
Grand Total
GeoXT
Post
1.44
1.44
SBAS
1.37
1.26
1.33
Uncorr
0.73
0.73
GeoXT
Total
1.34
1.26
1.32
ProXR
RTCM
0.38
0.61
0.49
ProXR
Total
0.38
0.61
0.49
Grand
Total
1.17
0.97
1.12
Location
Data
Count of
Feature Points and Positions
Corner
IP NW
Total Count of Point_ID
Total Sum of
Filt_Pos
Receiver
Correction
Count of Point_ID
Filt_Pos
Count of Point_ID
Filt_Pos
GeoXT
Post
9
1492
9
1492
SBAS
8
1280
5
905
13
2185
Uncorr
2
2836
2
2836
GeoXT
Total
19
5608
5
905
24
6513
ProXR
RTCM
4
2541
4
683
8
3224
ProXR
Total
4
2541
4
683
8
3224
Grand
Total
23
8149
9
1588
32
9737
Corrected test Points and separation from the antenna location.
<malbiblkchcpcpdh.png>
As above but with 2 uncorrected GeoXT points overlaid, including
the individual positions that were averaged.
<jpbjaanipeilbbgo.png>
Notes and findings:
Site is open sky but with house and trees adjacent
Antenna is static, occupation periods long (5 sec logging
interval)
32 observations averaged from 9737 positions
some observations are with the GeoXT internal antenna, others
are with a Trimble aircraft antenna (intended for SBAS)
Work flow as outlined above
The GeoXT uncorrected results are better than either of the
corrected results!?
The corrected ProXR results are better than any of the GeoXT
results, although biased to the east
The uncorrected GeoXT readings exhibit the NW bias but to a
lesser extent which seems to indicate that the correction does
not create the problem but may exacerbate it, if that makes any
sense.
I have probably missed something but my reaction remains that
the receiver may be defective (?)
Thanks again for your help and patience . . . . .
-----
Cheers, Spring
On 07/Mar/2021 03:54, kirk
wrote:
Hi Springer.
I assume your raw data files are being
converted to gpx on a computer since the raw terrasync
files are proprietary binary files. If you are using
trimble pathfinder, you can post process differentialy
correct the data if you have access to base station
logged at the same time you captured your field data.
Having a base station 100 miles away will not improve
your results as the baseline is too long.
I do not know if you can write a gpx file
directly from pathfinder but I would not bother. I would
write a shapefile which will contain the coordinate
system you specify. Simply open in qgis and you should
be good to go. If your older unit works better, I would
expect it may be an issue with the setup within
pathfinder or perhaps the software version.
I think your consistent offset is a direct
result of how you are converting your data from trimble
to gpx.
As I mentioned in my previous comments,
there are many issues which affect accuracy. Just
because the box says it is accurate you will rarely
replicate that in the field.
In terms of WAAS dataframes, these are
processed internally on your field unit.
Kirk Schmidt
Sent
from my Galaxy
-------- Original message --------
From: Springfield Harrison <stellargps at gmail.com>
Date: 2021-03-07 5:57 a.m. (GMT-04:00)
To: kirk <kirk at nortekresources.com>,
Jorge Gustavo Rocha <jgr at geomaster.pt>,
qgis-user at lists.osgeo.org,
Greg Troxel <gdt at lexort.com>,
Dan <19dmb83 at gmail.com>,
Nicolas Cadieux <njacadieux.gitlab at gmail.com>
Subject: Re: [Qgis-user] Trimble GeoXT 2005
Accuracy
Hello All,
Thanks for the comments, I'll reply more fully tomorrow.
The receiver is Trimble mapping grade:
"The GeoExplorer 2005 series consists of:
• The GeoXH™ handheld, providing subfoot (30 cm) accuracy, or
even 8-inch (20 cm) accuracy
with the optional Zephyr™ antenna.
• The GeoXT™ handheld offering submeter accuracy for GIS data
collection and data
maintenance.
• The GeoXM™ handheld with 1–3 meter GPS accuracy for mobile GIS
applications."
"Post processed carrier accuracy: 1-30cm". This receiver was
probably $5-8000 (?) new.
Data collection was stationary, open sky, good satellite
coverage, several minutes of 5 sec observations, good PDOP
SBAS and/or post processed
The concern is not the accuracy as such, but the systematic NW
shift. This has been observed over several months,
consistently. My old Trimble ProXR (1994?, $20K new!) is
actually better in this regard than the GeoXT!
The Trimble manuals make no mention of the SBAS CRS, implying
"turn it on and go, the receiver will integrate the SBAS into
the rover file."
More tomorrow, thanks . . . . .
-----
Cheers, Spring
On 06/Mar/2021 15:56, kirk
wrote:
A few notes.
sbas which is waas in north America is
based on equatorial satellites which will get you in
the 1 m range in southern Canada.
you can achieve sub decimeter accuracy
consistently using rtk ,either through a ntrip caster
(base station) broadcasting over the intenet or with
your own base station and a radio link. there are a
few chip sets and break out boards that you acquire
and assemble your own system. This is a very
inexpensive option.
Another option in Canada is to use
precise point positioning (PPP) which requires 6 to 12
hours of observation data using L1, L2 or L1 and L2
data and rinex log files. This comes in handy if you
need to establish a remote base station.
A proper antennae with a metal ground
plane is also critical to getting quality results.
Observing under a forested canopy is
difficult especially in summer under leaf on
conditions, after a rain which creates multiparth
mayhem.
There is a reason survey grade equipment
is relatively expensive. If you require repeatably
accurate results in a variety of conditions this is an
option.
Kirk Schmidt
Sent
from my Galaxy
-------- Original message --------
From: Jorge Gustavo Rocha <jgr at geomaster.pt>
Date: 2021-03-06 6:41 p.m. (GMT-04:00)
To: qgis-user at lists.osgeo.org
Subject: Re: [Qgis-user] Trimble GeoXT 2005
Accuracy
Hi,
I just jump in this thread to say I'm really impressed with
Ardusimple. I have a RTK Handheld Surveyor Kit [1] for +- 400
€ and it works really well.
I use the national NTRIP service and I have consistently
precisions around 10 cm with just one receiver.
I use a free Android application called SW Maps [2]. My
survey points, tracks and photos are collect in a geopackage
that I can read in QGIS. I use it mostly to collect ground
control points for my drone flights.
Regards,
Jorge Gustavo
[1] https://www.ardusimple.com/product/rtk-handheld-surveyor-kit/
[2] https://play.google.com/store/apps/details?id=np.com.softwel.swmaps
Às 20:12 de 06/03/21, Greg
Troxel escreveu:
Springfield Harrison <stellargps at gmail.com> writes:
Thanks Dan. See my relies to Kirk and Greg. The Emlid sounds
interesting, will have a look.
I have an earlier Emlid Reach (not RS or RS2), which has L1 only, and I
never got it to work well.
Also look at the Ardusimple unit -- but it's more a parts kit than a
system. You need a way to get RTK reference data in, and a good
antenna. One approach is Vespucci (OSM editor for Android) as a
datalogger, and the Ardusimple WiFi NTRIP master to get corrections over
the phone's hotspot.
https://www.ardusimple.com/product/simplertk2b/
_______________________________________________
Qgis-user mailing list
Qgis-user at lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-user
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
--
Email Signature
Geomaster
Jorge Gustavo Rocha
| Software
Engineer
e: jgr at geomaster.pt
| m: +351 910 333 888
g: 41.54094,-8.40490
| v:
510 906 109
a:
Rua
António Cândido Pinto, 67,
4715-400 Braga
_______________________________________________
Qgis-user mailing list
Qgis-user at lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-user
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20210309/f7a8dc15/attachment-0001.html>
More information about the Qgis-user
mailing list