Why the Pix4Dmapper-derived DSM values are negative?asking for more help

Hi guys,
      I have some UAV images, but unfortunately these images lost their longitude, latitude, elevation informations and POS informations.
      I still used Pix4Dmapper software to process these images to derive DOM and DSM. But I found the DSM values are negative, and the range of the DSM value are from -22.2872 to -20.8671.
      Now I have two questions, thanks to anyone who can give me the detail answers:

  1. The DSM values are negative. Can I say the unit of these DSM values was Metre?
  2. I want to know the reasons why the DSM values are negative? How does these negative DSM values been calculated in Pix4Dmapper? What is the principle of these calculation?

The following screenshots are about the DSM processing and its results.

Hi Guan,

Image geolocation and/or ground control points (GCPs) are used for georeferencing the project. Since this project does not have any initial reference information it is reconstructed in an arbitrary position. This is also the reason why the values for DSM are negative. 

In order to get the correct values for the results, I would recommend using GCPs in the project. More in the Using GCPs article. 

As a side note, since there is no geolocation in the project, it is highly possible that the orientation and the scale are wrong. It is possible to introduce scale and orientation constraints in order to resolve this:

Best,

1 Like

Hi,Blaž,thank you very much for your reply and your advise. Could you give more informations about my questions 1 and 2?

1. The DSM values are negative. Can I say the unit of these DSM values was Metre?

2. I noted that you say “Since this project does not have any initial reference information it is reconstructed in an arbitrary position”. But I tried many times under the condition of no georeference information, the DSM results are always negative value.  So I’m very interested in How does these negative DSM values been calculated in the Pix4Dmapper? What is the principle of these calculations?

Thanks a lot.

Hi,

When processing images with no image geolocation, and if no Ground Control Points are used, the final results are not scaled, rotated and oriented. Therefore they cannot be used for measurements, overlay, and comparison with previous results. This means that we cannot say whether the values you get in ArcGIS are meters or another order of magnitude. GCPs are required if there is a need for georeferenced outputs. In this case, the GCPs will scale, orient and position the final results. Additionally, they are very useful for increasing the relative accuracy of the outputs. 

A horizontal arbitrary coordinate system is a Cartesian plane that is not referenced with respect to a geographic system.
The position is defined by two linear coordinates: x and y. An arbitrary vertical system only defines a unit, but no reference surface so it is assumed that the horizontal system is also arbitrary and the two form an arbitrary 3D system.

Hello, I have a problem generating DSM. As in the previous query, it presents me with negative values, but the slope values are correct. The images captured by the drone are georeferenced and the mosaic generated is correctly located. What can be the reason?

Hi nicolaspaz,
Mapper will derive the values from the elevations recorded from the drone’s GNSS unit. Many consumer grade drones simply do not recorde the elevations very accurately. If you haven’t used GCPs then I would check the images first. But you can also upload your quality report to this and we can take a look.

Hello , the dsm and dtm are generated and all the results in the quality report are great but some point of the ground in the raycould is missing also the dsm result is not the same as the repport whats the problem please

Hi asma,
Can you attache your PDF quality report to this posting? We will take a look to see what might be going on.

Pfe2024_report.pdf (1.7 MB)

and a part of the orthophoto is missing


i hope you get back to me as soon as possible please , thank you

Hi asma,
I looked at your quality report and the problem has to do with your overlap. It is too low and that is why you are seeing these holes in the point cloud and also missing parts of the orthomosaic.

Lower-resolution drone images require 70-80% overlap. High-resolution PhaseOne images can get by with much less. But, in this project, it is still too low.

However, you might be able to improve the results with some changes to the processing settings. First, try increasing the image scale to 1. This will take much more time to process but it could help.


I would also try using the Alternative Calibration method.

I don’t know if this will help. But let us know.

thanks for answering , i tried increasing the image scale to 1 in the intial processing unfortianatly it didnt work , i even increased it to 2 also didnt work the gaps are still there , please can you give me any other solutions i need to do a vedio animation for this project and the holes aren’t helping and i need a great MNS .

And how can i know the résolution of my MNT

Hi @asma.daoud40,

You can find this in the Processing Options > Step 3. DSM orthoimosaic and index > Additional inputs.
Please share here a screenshot of the entire PIX4Dmapper window.

Daniele