Support Website Contact Support Blog

SOS!!! ADVICE NEEDED ! Photogrammetry Processing help for 10,000 image project !

Hello Pix4d Community;

I need help / advice and any suggestions!!

I am trying to process a 9,675 image (80 GB) project for 3D photogrammetry and ortho outputs. It is an existing pipeline corridor that is 24 miles long and 300 feet wide. The client wishes to view the ortho for erosion issues or other visually inspected issues on the pipeline, and the 3D outputs to make measurements. I flew with the Sensefly Ebee X, with the tilting SODA 3D camera, a PPK base and set 16 GCPs and will need to transform from WGS1984 to NAD83. It’s a fantastic dataset, that I am failing to process.

I am currently using a custom-build computer, that keeps crashing, or locking up. Here is my system:
Intel ® Core ™ i9-9980XE CPU @ 3.00 GHz
Samsung SSD 970 EVO Plus 1 TB (400 GB free) and I write to a SSD 3.5 TB drive
Video Card: NVIDIA GeForce RTX 2080 Ti

I am trying to meet my client’s need as soon as possible, but I am failing to complete sections. I have divided my project into 13 subprojects with about 750 images / sub, and this still fails half of the time. When it succeeds, about 7-10 hours. Does this sound right?

  1. Should I not be using Pix4d Mapper for this, or is my machine a wimp and I just need to make smaller Subprojects?
  2. Does recommend icloud Advanced for this task instead of local processing? I do have 16 GCPs and will need to transform the dataset from WGS84 to NAD83
  3. Should PixMatic be able to handle this size of dataset? If not, how many images would you recommend for subprojects? I downloaded a trial of Matic and am processing a subproject of 1,600 images

If anyone has advice on my Operation System or local vs cloud based results, any info would be greatly appreciated!

Thank you in advance,

Your computer specs are fine, but 64gb of memory really isn’t enough for this many photos if you want to process all at the same time. I’ve used Mapper and Matic. For a project this size I would absolutely use Matic instead of Mapper. It should easily handle your project. But definitely increase your memory to 128gb.

Hi @bhale,

I am more than happy to try and assist you with your project. We have quite a lot of experience processing large datasets, and corridor/pipeline projects.

Would you be willing to share your quality report’s? If so, please email to