Support Website Contact Support Blog

Capturing the pictures to my first project - How do I get it right from start?

I’m about to make a first trail on volume measurements for a potential customer. I’m used to drones, but not the details to capture the picture needed for the 3D model in Pix4D. I have read a lot about it and seen many videos so I got a rough idea about the process.

Goal: Provide a reliable volume measurement (one stock pile) for a potential future customer.

I’m planning to use my Phantom 3 PRO and the Pix4D app.

I will not use RTK GPS to determine the exact GPS location of GCP. Instead I’m planning to measure the exact distance between 3 GCP using laser. That should make it possible to set the scale of the 3D map correctly?

I will fly a double grid  pattern and use 80% overlap.

I’m not sure which height I should set. And I don’t know what camera angle that would be optimal for this kind of work. Any suggestions?

 

 

Hi Anders,

Since you will use Phantom 3 Pro for the image acquisition, we recommend to add GCPs in the project to correctly georeference it and correct for some inaccuracies that might arise in the Z direction.

Measuring the distance between 3 GCPs and adding a scale constraint can scale the project at the GCPs area. However, if the GCPs are concentrated at a specific region of the project the project will be locally scaled. Scale constraints should be added to the whole project area to correctly scale it.

The double grid mission is a special grid case that is designed for specific reconstruction purposes, such as facades reconstruction.
The camera angle can be adjusted between 45 and 80 degrees. These camera angles were tested and it was found that they give the best results. Setting the camera angle to 80 degrees will capture less background in the images and therefore you will have a better result for the 3D model. However, in case that you do not have very high stockpiles you can fly a simple grid mission without any difference in the results. 

The higher the flight height the more overlap between the images. The flight height will depend on the GSD you would like to achieve.
You can find more information about what is the GSD here: https://support.pix4d.com/hc/en-us/articles/202559809. And about how to calculate the flight height based on the desired GSD here: https://support.pix4d.com/hc/en-us/articles/202558849. 

Regards,

Thanks for the input. I flew the mission on a height of 50m. And double grid which resulted in a lot of images to process. I will revisit the site and try to map it again and fly from a higher altitude and use single grid. I will stay on 80% overlapping and camera angle at 80 degrees. 

I really wants to avoiding using GCP which are determined by a RTK GPS. Reason is that the needed equipment is very expensive. But I did measure the true distance in the area using a laser. Shouldn’t that be enough to set the correct scale of the project? I’m not looking for a exact GPS location for the map, I just wants to use it for measuring the volumes on stockpiles and shouldn’t  a correct scale of the area secure a  reliable measurement?

Another thing, being in Sweden we have a few months with snow on the ground. Of course I want to be able to do the measurements even if the stockpiles are covered in snow. The volume of the snow on the stockpile is fairly easy to manually calculate and remove from the total figure. But when I tried to map a stockpile cover in snow the desktop app had some difficulties to stitch the images correctly. I ended up with some huge black holes in the processed 3D map. The strange thing is that the 3D map which was processed in the cloud looked perfect. Any suggestions why the output from Cloud was perfect but not from desktop?

Hello,

You are right. Adding scale constraints in the model will improve its relative accuracy, especially at the region close to the scale constraints. Adding scale constraints around the stockpile should help improving the volume calculation. Please make sure that you reoptimize the project after adding the scale constraints. 

The problem with the snow is that it makes the reconstruction of the area more difficult. Snow has little visual content due to large uniform areas. Therefore, a higher overlap (at least 85% frontal and 70% side overlap) should be used and setting the exposure settings accordingly to get as much contrast as possible in each image. 

Which version of the software are you using? On the cloud the latest version is always used. 

Regards, 

 

Im using version 3.0.13 (running on Mac)

100m and 90% and I manage the reconstruction of the areas looks perfect. And thats with stockpiles totally covered in snow. Im pleased :slight_smile:

 

“…adding scale constraints in the model will improve its relative accuracy, especially at the region close to the scale constraints. Adding scale constraints around the stockpile should help improving the volume calculation.”

Using scale constrains, will this also improve the accuracy on the elevation of the stock pile?

 

 

Hi Anders,

Yes, scale constraints will also improve the relative accuracy in elevation of the stockpile. For more information: https://support.pix4d.com/hc/en-us/articles/205360375

Best regards,