next is my log，what should I do to solve it？
[UI]: New Project clicked.
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Processing]: ProjectLoading
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Processing]: ProjectLoaded
[UI]: Processing Start clicked.
[Processing]: Step1Running
[Warning]: Warning w9020: GDAL Error \<1\> latitude or longitude exceeded limits
[Processing]: Substep Keypoint extraction started.
[Processing]: Gather image information
[Processing]: Generate normalized/thumbnail images
[Processing]: Computing keypoints
[UI]: Processing Options clicked.
[Warning]: Warning w9021a: DataSender - Network connection problem.
[Warning]: Warning w9021a: DataSender - Network connection problem.
[Processing]: Write Keypoints.
[Processing]: Substep Keypoint extraction finished.
[Processing]: Substep Keypoint matching started.
[Processing]: Generating pairs
[Processing]: Computing matches
[Processing]: Write matches.
[Processing]: Write matches.
[Processing]: Substep Keypoint matching finished.
[Processing]: Initialize automatic tie points.
[Processing]: Building automatic tie points.
[Processing]: Substep Camera calibration started.
[Processing]: Analyse automatic tie points.
[Processing]: Analyse automatic tie points.
[Processing]: Compute number of matches.
[Processing]: Initialize calibration
[Processing]: Initialize calibration
[Processing]: Calibrate
[Processing]: Initialize calibration
[Processing]: Calibrate
[Processing]: Initialize calibration
[Processing]: Calibrate
[Processing]: Initialize calibration
[Processing]: Calibrate
[Processing]: Initialize calibration
[Processing]: Calibrate
[Processing]: Initialize calibration
[Warning]: No initial frame
[Processing]: Initialize calibration
[Processing]: Initialize calibration
[Warning]: No initial frame
[Processing]: Transform init.
[Processing]: Substep Camera calibration finished.
[Processing]: Add cameras...
[Processing]: Substep Rematch started.
[Processing]: Substep Georeference the model started.
[Processing]: Optimizing with geoinformation
[Processing]: Analyzing geoinformation
[Processing]: Check consistency.
[Processing]: Compute number of matches.
[Processing]: Substep Georeference the model finished.
[Processing]: Substep Rematch finished.
[Processing]: Substep Camera calibration started.
**[Error]: No calibrated cameras.**
[Processing]: Substep Camera calibration finished.
[Processing]: Substep Report generation started.
[Processing]: Read keypoints.
[Processing]: Substep Report generation finished.
**[Error]: Report processing: invalid map\<K, T\> key!**
**[Error]: Error e0046: Processing failed. No calibrated cameras.**
[Processing]: ProjectLoaded
[Warning]: Warning w9021a: DataSender - Network connection problem.
[Warning]: Warning w9021a: DataSender - Network connection problem.
[Warning]: Warning w9021a: DataSender - Network connection problem.
[Warning]: Warning w9021a: DataSender - Network connection problem.
[UI]: New Project clicked.
[UI]: Processing Options clicked.
[UI]: Processing Options clicked.
[UI]: New Project clicked.
Hi Yang,
If no cameras are calibrated then Step 1 will fail and processing subsequent steps will not be possible.
This could be caused by:
- Image quality is poor (blurry images, overexposed, etc.) or there is too little overlap between the images.
- Recapture images that are clear and properly exposed. How to verify that the Image Quality is sufficient
- Verify that there is enough overlap between the images. How to verify that there is enough overlap between the images.
- The image content is not appropriate for photogrammetry due to low texture (water, glass, sky, etc.).
- Recapture dataset so that individual images contain at least 40% content that is suitable for photogrammetry.
- The camera model used for the project is not appropriate for the camera used to capture the image dataset.
- Verify that the camera model associated with the camera is correct. For more information: How to verify that the Camera Model associated to the Camera is Correct.
Hope this will help.
Best,
1 Like
Hi Marco,
i have the same problem. see image attached
In my project I separate the image sequence to strips, for example: I have 14 strips and I do mosaic for strips 1-3,4-6,7-10 and 11-14. The bug I encounter is not repeated for all strips but for some, meaning, the pix4d run successfully on strips 1-10 and failed on strips 11-14. Image quality, overlapping and image texture are the same for all strips, therefore I don’t find those reasons as main reasons for failure. Can you think of something we can try in order to make mosaic also for the strips that failed?
Regards,
Hadar
Hi @gilboa.hadar,
What is the overlap between your strips?
Are they all presented in your screenshot?
Could you please attach here the quality report and the log file of your project?
From this, we might be able to recommend you some processing options that could make it work.
Best,
Dear Marco,
Thank you for the prompt response.
See attached two reports attached. One for Nov 14 strip 1-4 which was successful run and the other for nov14 strip 11-14 which failed.
The overlaps between strips was 65%. Unfortunately until now from 8 flights we have only one complete mosaic while in the rest we have only partly mosaics. The flights was done with DJI Matrice 600 and Flir A655sc thermal camera. Before we place images in the pix4D we gave each image from the Flir the relevant geodata from the drone.
Regards,
Hadar
(Attachment report_14Nov19_strip1_4.zip is missing)
report Nov14_noon_strip 11-14.zip (90.9 KB)
Hi Marco,
I got rejection message on one zip file I sent. Here is only the xml report of the successful run.
Regards,
Hadar
report.xml (18.3 KB)
Hi,
I guess that your project did not generated a quality report in the *.pdf format? Only the *.xml right?
Do you have the log file?
You can drop the files here if attaching files does not work:
Google Drive
But I can already say few things.
If you want successfully process thermal project, the overlap should be at least 90% side and front:
Then you is your camera defined? Did you edit your camera model?
Attached you will find the camera specifications.
FLIR A655sc 25°.pdf (30.6 KB)
You can also add the geolocation for your images in Pix4D.
Best
Hi Marco,
Thanks for your response. We work in 65% side overlap because it works for us in a former project and since the feild we test is quite big (380 Dunam) we decided to keep working with the 65% in order to have minimum temperature differences along the whole flight. We already did 8 flights with the same flight parameters (11.8m/s drone speed, 65% side overlap, 94% horizontal overlap). Because we already have all data, and the feild is now already cover with vegetation we can’t repeat the flights, therefore we need to do our best with the avilable images.
I wish to send you two logs, one of succesful mosaic and the other that failed. I try to upload it to the link you sent me but it’s lock. Can you please check it and send it again.
Thanks,
Hadar
Hi Hadar,
I cannot guarantee that your data set will work, but we can give a try.
I am mostly interested to see which processing options have you been applying.
In your case, it might be useful to work with the different strips in different projects and then merge them.
The link should now work. Sorry about that.
Best
Hi Marco,
I used the default processing except two changes, in the intial processing I changed the keypoint image scale to: custom and checked the double image size. In the point and cloud option I checked the generate 3D textured mesh (with the defaulf FBX). I didn’t changed any other processing paramters and I did the same for all strips.
Regards,
Hadar
Hi Hadar,
From your logs I have only been able to detect an error during the calibration that I cannot explain.
My first guess would be that there is not enough overlap.
The only thing I could recommend you to try but I do not guarantee that the results will be better.
Set the calibration method on “Standard” and the Internal Parameter Optimization on “All Prior”.
Usually, the “Alternative” pipeline, the one you use, if recommended in flat area such as yours.
But I would give a try as I don’t see many other options.
For further investigation, we will need to test your data set on our side.
If it does not improve at all, could you please upload your images on the Google Drive shared above?
Thank you!
Best
Hi Marco,
Thanks for your response. I just saw it now after I tried a different calibration approach, that seem to work. Instead of “Alternative” I used the “Accurate Geolocation and Orientation”. So far ,with the new calibration method, the SW successfully complete the mosaics for areas that failed with the “Alternative” method. I will also try your suggestion with the “Standard” method. If both methods will work, which method in your opinion is the correct method for this kind of processing?
Thanks,
Hadar
Hi Hadar, Accurate geolocation and orientation works when you have orienation data and also the geolocation and orienation values should be accurate (like RTK). But from our exprience we have seen thsi method also working for normal cameras. When you have nadir and oblique imagery both, this option works the best. Also, for homogeneous areas, it would work better than standard. Standard calibration is used when you have a lot of heterogeneity in terms of texture, object height (like buildings etc). Alternative calibration is used for homogeneous areas but is based on the assumption that the images are nadir and the area is flat. If the angle >35% (sue to oblque camera angle, slope) Pix4D fails to calibrate such images with alternative calibration. Thus in these cases, accurate geolocation and orientation works better
good afternoon, I am trying to use pix4d to realize dtm of diverse area of ganimede planet for structural analysis of the planet.
I am trying to realize dtms from orbital images, but the software gives diverse errors, someone can help me?