Pix4D Desktop, Wrong Computed Camera Position.

I was having some problems with an error with writing the exif file. Support directed me to download the newest version which I did. So the program processed from scratch and when I got back to work this morning to check on the results I notice a problem.For some reason it has calculated some of the camera positions to be way higher than they really were and has shifted them to weird angles causing a giant blob to come up out of the ground in my point cloud. I used Pix4d to control my DJI M210. I have 11 GCPs. They all look great except for the one near this anomaly. The computed position and initial position do not make any sense. Computed position is way up in the air by about 200’ and if you check the images where it marks the control point you can see it completely missed the control point and just marked some random spot on the ground. I’ll include a couple screen grabs for reference. Anyone know how I can fix this problem? I need to be able to deliver this product soon.

I’m very new to Pix4D so sorry if I am missing something obvious. I’m a professional Land surveyor so I know my control points are good. I don’t know what the problem could be. Any advice is greatly appreciated.

 

Have you checked if the CGP that is outside the tolerance and near this artifact has the wright coordinates imputed?

Yes. The coordinates appear to be correct.

Hello Joshua,

Could you please check the quality of the images in this area? Are all of them of high quality? 

Could you please remove this GCP and Reoptimize the project? How does the model look like now?

Regards,

So I removed the GCP and re-optimized steps 2 and 3. and it still appears as if it did not make a difference. Please advise with the next thing I should try.

anyone? I could use some help.

Any chance you could share the dataset? I could take a look at it

Hi Joshua,

Could you please send me the dataset (.p4d file, images, GCPs file) to better understand the cause of the issue?

You can use any file sharing service to share the files (Dropbox, Google Drive, WeTransfer).

Regards,

Hi Joshua, 

I would like to help, like Antigoni, just send the dataset to me and I will check it out.

We had a similar experience recently. We started a new project. We marked the target in only three images for each GCP then re-optimized and marked the remainder of the photos. That worked for us.

I am also having similar issues starting two weeks ago using Phantom-4 drone and an old version of pix4d desktop since my support package ended

but im getting exact same issue and sometimes it creates2 intersecting planes 1 properly oriented and one tilted and intersecting with the first i checked the images exif nothing seemed out of place and the images also seems correct

in some projects reprocessing them with optimize using only geocoordinates seems to resolve the problem

Hello all,

Every project is a different case. Please share here some screenshots of the rayCloud and the PDF quality report of the project.
Are the images taken during different flights? Is this a merged project?

In the meantime, try to add Manual Tie Points in the intersection of the two planes and in the rest of the project area where you observe that there is a low amount of matches.

Regards,

hello all,

i want to know. why this problem will occur and how to solve…

Hi Atul,

We would need more details of your project to provide assistance in the most efficient way. Could you please send us the following?

▸ The quality report (.pdf format): …\project_name\1_initial\report\project_name_report.pdf
▸ The project log file (text file): …\project_name\project_name.log
▸ The project file (.p4d): …\project_name.p4d

To store the files you can use our OneDrive

Thanks in advance!