Messages

1

General / Re: Relation between markers' accuracy and error

« on: March 21, 2023, 11:43:16 AM »

Rok,

program is using gauss equalization, therefor you get the mean deviation. Absolut deviation can be much higher.

regards
Martin

2

General / Re: Aligning photos, optimizing cameras and gcps - HELP PLEASE

« on: March 21, 2023, 11:37:18 AM »

Shovelhead,

as a start,  I prefer Twincam :-)

You can adjust the settings before you import the cameras, if you want to use the rtk position of the pictures or not.  (Load camera location acurracy from xmp meta data).
If I refuse  that, the accuracy between the  gcps increases (down to approx. 1cm).
But this shows just  the deviation of the location between the gcps, not the accuracy of the model.

regards
Martin

3

General / Re: DEM Transformation AUSTRIA for RTK Modells

« on: March 21, 2023, 11:05:10 AM »

good morning Kiesel,

yes, that works good. But then, the money spent for the RTK Mode of the Mavic 3 E was frustrated.
My optimum would be using the RTK Mode, having 3 - 5 GCPs (usally i use 2-3 pcs per Hektar), calculate the deviation and transform the modell (using x,y,z movement, if the area is not too big)

that would have been the easiest way.
Unfortunately i have not found a possibility to do so in Agisoft and QGIS.
iT WORKS on Bricscad, but  the way is alittle bit complicated.

regards
Martin

4

General / DEM Transformation AUSTRIA for RTK Modells

« on: March 18, 2023, 09:42:51 AM »

Good Morning,

I am using Agisoft since 7 years now. First with an German Multikopter, the a Phantom 4 Pro and now a Mavic 3 Enterprise and RTK modul.

Main purpose is surveying of quarries.

Since the beginning I have a Trimble R6 Rover, using the austrian apos correktion data. Works pretty well.
The austrian Grid (EPSG: MGI 31255 or 31256) need a local correction on the trimble rover. That means, Fixpoint are measured and there has to be a local correction on the survey controller.

This can not be applied on the Mavic RTK system. Therefor there is a deviation between the GCPs (measured by Trimble) and the RTK Model.

Difference can bei:
X = -0,02 m
Y = 0,64 m
Z = 2,34 m (actual values from this week).

Now the question:
- If I use the GCPs on the RTK Modell, the overall deviation is about 0,10 m, but the modell difference on height ist stll to much.

- Is there a funktion to get the DEM and Ortho modified by these values?

regards
Martin

5

General / Re: Using RTK and Ground control Points

« on: March 04, 2023, 07:01:43 PM »

Hello,

I have found the item in the general settings. No caculating again, will see.

regards
Martin

6

General / Re: Relation between markers' accuracy and error

« on: March 04, 2023, 06:20:02 PM »

Hi,

using the trimble R 6 Rover the accuracy between the markers is about 0,015 cm.

Regards

Martin

7

General / Using RTK and Ground control Points

« on: March 04, 2023, 05:56:06 PM »

Hello Together,

I am using Metashape since 2015 for surveying of quarries. I love the programm, the results are very good.
Usually I use a Phantom 4 Pro and a lot of ground control points (up to 3-4 / 10.000 m²). GCPs are measured with Trimble R6.

Now i have bought a Mavic 3 Enterprise, using also the RTK Modul (using Network). I works well, BUT:

Total accuracy is lower than using just the P4P. This is, because Trimble R6 is using a local ajustment to the austrian reference system (which is not available to the MAVIC).

Total error is about 0,08 cm, which is not so much but gives huge deviations in VolumeCalculation.

Now the question: How can I decide, using the "location of the fotos or not?"

Thx,

Martin

8

General / Re: Significant improval between Metashape 1.7. and Agisoft Photoscan Prof. 1.2.6.

« on: February 18, 2022, 01:23:43 PM »

Alexey,
downloaded the last version Metashape 1.8.
Unlocking not working.

Any Ideo what I have done wrong?

Regards Martin

9

General / Re: Significant improval between Metashape 1.7. and Agisoft Photoscan Prof. 1.2.6.

« on: February 18, 2022, 01:03:56 PM »

Hello Alexey,

usually I use the error on the GCPs, sometimes I check also with Check points.
I will try the 1.x.x Version of Metashape.

Thanks

Martin

10

General / Significant improval between Metashape 1.7. and Agisoft Photoscan Prof. 1.2.6.

« on: February 18, 2022, 10:11:20 AM »

Hello,

I am using Agisoft Professional 1.2.6 2843 (64 bit) since 2016. I am satisfied with the results.
I am currently using a DJI Phantom 4 Pro (fixed shutter, 20 MP, Camera uncalibrated) for surveying purpose.
Usually flying at an altitude of 100 m (85% Front overlap, 67 % Side overlap)  I am getting an error of 0,185 (pix) and a Error of 0,027m of the GCPs.
I am using usually 15 - 30 GCPs on an area of 7 - 20 hectares, using a Trimble RTK GPS System (inkluding Glonass) for GCP Measurement.

Absolutely ok, no more accuracy needed. Ground solution is about 9,4 cm/pix; point density 113 pts/m².

Now I have tried a photogrammetry of Fotos made by DJI Mavic 2 Pro. Camera uncalibrated; The results are not well. It can depend on low number of GCPs (only 8 of them on 25 Hektares) and high number of fotos (475 #).

Best result was: 0,214 Image error; 0,077 m Total error. (using only 50 % of the pictures). Reprojection Error was 4,22 pix.

The best result of Metashape Professional 1.7.2. was total error xyz 3,71 mm (!), which is at least half of the error.
Camera Reprojection error was 7,02 Pix.

Questions:

- is the algorythm of Metashape that better than the old version of Agisoft?
- Is there any improval for analysing rolling shutter fotos?
- is ist worth spending money for an upgrade?

Regards

Martin

11

General / Re: PhotoScan to AutoCAD best workflow

« on: March 13, 2017, 09:11:57 PM »

first of all export dem and ortho as a tiff, using the grid as you need.
Then drawing the edge lines in Qgis using shadings of the dem.
you are getting a shape file with the levelling 0.
I then use a self made program to export the shape file to the dem and generating the dedges as 3d polylines as a dxf.
import the dxf to Autocad, or even better, to Bricscad.

For the purpose of modelling, i use the isolines generated in qgis or calculate a raster in qgis, using point sampling tool to make it 3d.

there are cheap programs for the shape and dxf conversion if you have no civil 3d, whicht is much to expensive ans autodesk are no friendly people at all.

regards
Martin

12

General / Using QGIS for Control of the DEM; Surveying of Quarries

« on: March 08, 2017, 12:16:29 PM »

Hi Folks,

I am always looking on the accuracy of the DEM and possibilities to test it. I am using GCPs, measure them with GPS (accurate Trimble one).

Now the idea is: I measure different points on the surface of the quarry (between the GCPs) and around the model. I do not use any markers on this points. I export the measured points into a shp-file (as points) and use the point sampling tool of qgis to make a projection on the surface of the DEM.
Export the points into CSV-File and compare the surface levels with the GPS - measured heights. 

Advantage: information from outside AGISOFT and quicker than checkpoints.
Disadvantage: 10 min additional work on the PC. No information in x / y deviation.

Any Ideas and comments?

Regards

Martin

13

General / Re: Vertical Misalignment with local control GCP's

« on: March 02, 2017, 08:11:13 AM »

Hi Sav,

i am doing that report. Is it possible to export the overlapping map to qgis?
I am still meaning the bulging of the model outside of the gcps when overlapping is still ok.
any ideas on that?

regards
Martin

14

General / Re: Measuring area

« on: February 28, 2017, 06:41:30 PM »

use qgis to calculate a meshand use this for further calculation.

regards Martin

15

General / Re: Vertical Misalignment with local control GCP's

« on: February 28, 2017, 12:05:01 PM »

Hello,

I am usually using high accuracy GPS RTK (Trimble) for the GCPs. I am using no GPS information for the camera positions.
LAst time I recognized, that the height information is getting inaccurate to the edge of the model.

Are there any recommendations, how wide the gcps should be spreaded to the egde of the measurement?

Are there any rates of bulging of the model?

Regards

Martin