When entering GCP’s, it is recommended to include their accuracy in meters. We are using GNSS Real-time SBAS Correction with Trimble Pathfinder software and a Geo 7X (H-Star) unit. Our metadata provides HDOP and PDOP for accuracy estimation, which only includes one component of positional error (see https://www.gps-forums.com/threads/roughtly-converting-dop-to-metric-error.40105/, https://www.trimble.com/OEM_ReceiverHelp/V4.44/en/NMEA-0183messages_MessageOverview.html) and is unitless, I believe. However it provides horizontal and vertical precision in meters. While the Pix4D documentation says it wants accuracy, does it actually want precision instead? If it wants accuracy, how to we obtain accuracy in meters from a Trimble data file for each GCP individually? Or are we supposed to enter the same accuracy for all GCP’s collected from the same GPS model?
It means precision in meters. That’s what the accuracy means in the GCP menu
Normally I let it be with the standard 0.02 for RTK GPS positioning of my GCP
Hi,
As Jorge already mentioned, you can start by setting the accuracy to 0.02m in the GCP/MTP Manager. Typically this is the value that is used in most of the project with GCPs.
After Step 1 is processed you can check, in the quality report, if the calculated errors correspond to the estimated accuracy.
More here: Quality report specifications - Ground Control Points
Best,
Thanks, Jorge and Blaz! Blaz, are Jorge and I correct in thinking that the “accuracy” field actually refers to precision? If so, I will enter the precision that the unit already estimated, which varies depending on the point from 0.4-1.4 m. If you don’t mean precision, I will use 0.02 for all points.
And is 0.02m typical for vertical accuracy as well as horizontal? If not, what is a good value for vertical?
I only noticed now that you are talking about the SBAS Correction which, indeed, has the position errors around 1 meter. The example that we mentioned with Jorge applies for RTK units with the precision of few centimeters.
The ideal case would be that you have the precision estimate for every point in meters that correspond to the real conditions at the time the point is measured (number and position of the satellites, …). Maybe some other users in the community know how to convert the values in the metric system?
As you do not have metric values at the moment, I would recommend processing the project with the estimated values for precision in the GCP/MTP manager as shown below:
Best,
Okay, so we are correct that you mean precision, not accuracy? The difference is illustrated here: http://cdn.antarcticglaciers.org/wp-content/uploads/2013/11/precision_accuracy.png. I have the meters of precision for each individual point, but I don’t have a point-specific estimate of accuracy, just the following general estimate.
The metadata from our GCP’s shows receiver type: “Geo 7X (H-Star)” but correction type: “Real-time SBAS Corrected”. The specs for our unit are at http://compasstoolsinc.com/wp-content/uploads/022516-002E_Trimble-Geo7_DS_USL_0216_LR.pdf, showing “H-Star™ accuracy (real-time or postprocessed):10 cm + 1 ppm (=mm error/km estimated) HRMS” and “SBAS accuracy: <100 cm”. I don’t know how to calculate HRMS given then I only have PDOP and HDOP, no estimates for other sources of error (https://www.gps-forums.com/threads/roughtly-converting-dop-to-metric-error.40105/). .%C2%A0) I am interpreting these specs to mean that the H-Star technology improves the SBAS accuracy from the <100cm that it would normally be to now 10 cm + 1 ppm HRMS. If the project is 3 km across, then 1 ppm = 3 mm, so effectively my horizontal accuracy would be approximately 10.3 cm? If accuracy is what you are looking for, not precision, I will enter this for all points.
Yes, the correct term is precision. I agree that the term we use in the software can be confusing and I informed the team in charge in order to consider changing it.
Actually, the accuracy is estimated in the Pix4D software and can be seen in the Quality report after the GCPs are incorporated and the project, more specifically Step 1., is processed. After the processing is done the inputted precision of the GCPs is one of the parameters used to calculate the estimated accuracy.
In case that you need further information about the configuration of the system, correction type that was applied and actual values that the system provides, I would recommend you to directly contact Trimble support.
Wonderful, Blaz, thanks so much for clarifying and it would be great if you could also correct the software and documentation like you proposed!
Hello, may I ask whether setting the horizontal and vertical accuracy of ground control points will affect the accuracy of 3d model?Looking forward to your reply.Thank you very much.
I am still a bit confused on the terminology Pix4dmapper uses regarding the “Accuracy” of the input GCPs. The default values of .02m for the Z is representative of an RTK dual-frequency multi-constellation receiver if the value is reported in RMSE (68% confidence value). The term accuracy also makes sense if the algorithm weights the GCP more important if the “accuracy” value is smaller. Either way the confidence % (e.g. 63, 68, 95) must be included whenever using the term accuracy. Accuracy can be inferred from other empirical work or estimated through error budget modeling. For example, Topcon’s data collectors estimate accuracy based on the observed constellation (also emulated in the PDOP and DOP sisters metrics) and occupation averaging time.
The term precision, instead of accuracy, does not make sense with the examples provided. Precision typically refers to the number of significant digits (e.g. meter, decimeter, centimeter) in the position that are reliable or measurable. Would the default .02m mean the GCP values are only reliable in 2cm increments?