poor quality orthomosaic quality

PIX4Dmapper PIX4Dmapper Questions/Troubleshooting
1

I have produced an orthomosaic in Pix4D mapper from a DJI P4RTK flight over a crop and would like to create a TGI index map from it but the ortho has many artefacts. I also ran the images through an online photogrammetry portal and the ortho is perfect. Can you tell me how to change the parameters in Pix4d Mapper to get the best possible orthomosaic quality.

Cheers,

2

Hello,

could you please upload the quality report of your project that can be found here (.pdf format): …\project_name\1_initial\report\project_name_report.pdf
and some screenshots that highlight the issue?

3

Hi Christina,

Thank you for your prompt reply to my query. Attached is the Pix4D Mapper report.

image001.png

Bundy Research Station Legume Fallow Trial P4RTK 17-02-2021_report.pdf (2.03 MB)

PMN Trial Layout 2.jpg
PMN Trial Layout 2.jpg1817×731 779 KB

PMN Trial NDVI Pix4D Zoom.jpg
PMN Trial NDVI Pix4D Zoom.jpg1735×743 413 KB

(Attachment Bundy PMN Trial.docx is missing)

(Attachment DAF Bundy PMN Trial Index Calculator Values (version 1).xlsb.xlsx is missing)

4

Hi Christina,

Some of the attachments to my last email were not accepted by your system so I have uploaded them to my OneDrive.

image004.pnghttps://itpqld-my.sharepoint.com/:f:/r/personal/marcus_bulstrode_daf_qld_gov_au/Documents/Pix4D%20Mapper%20Information?csf=1&web=1&e=WaNNfM

Cheers,

image001.png

5

Hello Marcus,
The artifacts can be minimized with good dataset. You could generally increase your flight height and the image overlap to minimize the effect. For your quality report, I don’t see any issue while processing. I would suggest you to try the below calibration methods and see if that is helpful.

  1. Calibration Method: Alternative (Optimized for aerial nadir images with accurate geolocation, low texture content and relatively flat terrain, fields for example)
  2. Internal Parameters Optimization: All Prior (Forces the optimal internal parameters to be close to the initial values.)
  3. Increase the image scale to full during step 2. Point, Cloud and Mesh.

I would also recommend you to go through our support article on Distortions and Artifacts in the Orthomosaic.

6

Hi Kapil,
This is the first opportunity that I’ve had to get back to this project. There are a couple of things that I don’t understand. I am running it now with my interpretation on the parameter that you suggested but i am unsure if I have interoperated them correctly.

1. Calibration Method: Alternative (Optimized for aerial nadir images with accurate geolocation, low texture content and relatively flat terrain, fields for example)

Where do I find the “Alternative” option. Is this AG RGB?

2. Internal Parameters Optimization: All Prior (Forces the optimal internal parameters to be close to the initial values.)

Where is this option located?

  1. Increase the image scale to full during step 2. Point, Cloud and Mesh.

Is this “Original image size, Slow” ?

Cheers,
Marcus

7

Hi Kapil,

These are two other questions that I sent through with my support query. Are you able to answer these for me?

  1. Whilst I have you, I have some other questions. In a different project (see attached PMN Layout) I am determining the NDVI values for individual plots across a trial (see attached PMN Trial NDVI). So far I have used the Area % value and this has produced some useful statistics (see attached excel). What I would like to do is use the same statistics (see attached Bundy PMN Trial.docx) using the Area value or, if I can the pixel number. The current report only shows area as “ha” but I need more decimal places. For example, instead of 0.01, I need 0.0676. Is there a way that I can obtain more decimal places and/or the actual pixel count for each colour?

  2. As I mentioned, I have a DJI P4 Multispectral drone and the new MicaSense calibration reflectance panel (previously I used a Sequoia and older style panel……both were substandard). When I load a DJI P4M image set into Pix4D Mapper, does the software recognise the EXIF data and account for irradiance and the calibration values from the target? Also, if I load a DJI P4M image set into Pix4D Mapper without the reflectance target images does the software still recognise the EXIF irradiance values?

Cheers,
Marcus

8

Hello Marcus, If the calibration option is not available, you have to check the Advanced option as shown in the figure below. Higher the image scale higher the processing time.

image
image802×732 31.7 KB

9

Hello Marcus,

  1. Regarding the area in decimal and the pixel number, you should be able to do an area measurement. For more information, visit How to measure Distances, Areas or Volumes. However, I believe that using post-processing software such as QGIS or ArcGIS would be more efficient.
  2. The calibration panel, I would suggest you to go through the below support article, Radiometric calibration target
10

Thanks Kapil,

I have been using the software for years but never seen the Advanced check box before! I am running the data set again now.

Cheers,

image001.png

11

Hi Kapil,

Here is a OneDrive link to what I am trying to do.

image004.pngPix4D Mapper Information

You will see in PMN Trial NDVI Pix4D Zoom.jpg that the Index Calculator report only has 2 decimal points. This means that on a small research plot the Area value is not accurate enough to show the difference in a meaningful way (0.00 or 0.01).

Are you suggesting that Pix4D Mapper does not have any finer level settings and that I will need to do a raster assessment in another platform such as QGIS?

Also how can I process all of my plots at one time instead of having to do each one individually?

Cheers,

image001.png