Need help with my first project with Pix4D

Hi

My company have bought 1 month subscription to really test if Pix4D can deliver what we need.

I was playing around with the software during the 10 day trial, but couldn’t really get it to work properly and defaulted back to Agisoft Photoscan.

Now, since we have payed for 1 month sub we are really going to test it.

Information about the project:

  • Areal mapping, consisting of 2 flights with 283 images.

  • I have managed to merge the datasets

  • I have imported GCPs with x,y,z from a high accuracy GPS.

Now, my first issue:

The GCPs looks like they are underneath the dataset, how do I correct the set to use the GCPs?

Hoping for a swift reply

 

Marius 

 

Marius,

This is not un-common.  

All you need to do is Mark 3 GCP’s (w/ 3 images each, min) in the Basic Editor, apply, and then your model will be shifted down to match your GCP vertical coordinate system (from GPS).  Then you can fine-tune your GCP marks in the Raycloud Editor.

 

 

Hey, thanks for your input.

Yeah, I figured it out after watching one of the webinars. The altitude accuracy on a phantom 3 is horrible.

 

Anyway, my next hurdle is the export of dense point cloud.
Problem: I wanted to see what Pix4D could push out so I had everything set to max. 1 day later it gave me a .LAS with 15 parts.

Is there any option to force it to make 1 big one??

 

Marius

Figured that one out aswell, so never mind.

 

Another problem:

One of the missions failed mid-flight. I restarted it, and it completed. Problem is that it only synced the failed mission to my device.
I have confirmed that the images from the successful flight are indeed on the drone.

Anyone experienced this, and how did you solve it?

Marius,

You can still process the images w/ Pix4D.  

Just start a new project and import the photo’s directly. 

Your images still have geolocation data in the EXIF headers (although this is not as precise as using the .p4d file produced from the CAPTURE APP).  But since you are using GCP’s, you shouldn’t notice a difference except possibly slightly longer processing time.

 

On large projects, I’ve flown multiple missions with Pix4D CAPTURE APP to get the required coverage/overlap, but didn’t use the .p4d files CAPTURE creates.  I just created a new Pix4D project and used all the images (from SD Card) collected in a single Model.   

The 2 most important things I’ve learned are to make sure you have GCP’s outside the boundary of your project (not just in the interior of your area), and to add Manual Tie Points.  MTP’s have a huge impact on the 3D Reconstruction.  

 

 

 

Hi Marius,

A synchronization problem might happen at the end of the first mission which created issues on the second mission that was successful.

It is still possible to create a Pix4D project from scratch using the images on the drone’s SD card.
This article will probably help: https://support.pix4d.com/hc/en-us/articles/204911905

Processing from the .p4d file generated by the app is recommended but not necessary.

Regards,

Wouldn’t it be off when you don’t use pix4d data file? The image Metadata doesn’t contain yaw, pitch and roll.
Marius

Our software Pix4Dmapper does not need the yaw, pitch, roll information. If by metadata you mean EXIF of the images, then with the latest DJI firmware of certain drones they now measure and record the orientation of the camera besides the position.

Processing with or without the .p4d is the same in terms of results. You should not worry about it :).

Regards,

I never use the create .p4d files. Im out in the field with no internet, nor desire to wait to turn off my drone. So as soon as I land, i X out the synchronising bar and turn everything off. Then once back in the office, I create a folder on my Pix4d drive and cut and paste the pictures to it.