Difficulty Getting a Successful Processing Solution for Forested Area in Pix4dMatic

I am experiencing trouble trying to process a forested area with only partial solutions using a trial version of Pix4dMatic. Currently I am conducting a “Pillsbury Bake Off” between Pix4dMatic and a competitor to decide which software application will be my go to platform in the future. The project in question is a 2600 acre area that is primarily logged forest in Northern New York that was flown in April 2020 for testing and evaluation purposes on my hunting club’s former lease property performed during the Covid lockdown period (What a great way to social distance during that period I might add). The competitor’s application does not have a problem processing the same area. I have tried various combinations of Calibration exhausting all combinations of Image Scale (1/1, 1/2, 1/4, 1/8) Keypoints (Auto as well as custom (2000000, 1000000, 500000, 50000, 10000, 5000, 1000, etc) and Low/High Internals Confidence all to no avail. Here are the flight parameters. DJI Phantom 4 Pro flown at 400’ AGL with a 75% Overlap and a 65% Endlap. Now I know that I should have increased that combination to something like 85%/80% but I planned and executed this at the last moment and had to fly everything before the leaves popped on the trees. Here is a sample area ortho of one of my flight plans for the unsuccessful Pix4dMatic ortho and the successful competitors ortho. If anyone would like to take a stab at this dataset you can access the raw imagery here: Amazon Drive
Any processing suggestions other than to re-fly (not an option as the other application currently processes the dataset in it’s current state and would probably be a deal breaker for me if that was the only solution) the project at 85/80 would be greatly appreciated. Thanks.

Pix4dMatic Unsuccessful Processed Ortho (All variations of Calibration described above produced similar unsuccessful results.)

Cameras

Competitors Successful Ortho:

Hi @DH_79 ,

You reached out to us also via personal email and we have responded to you there.

Best,

Hi David,

We had a look at your dataset and we managed to process it
in PIX4Dmapper by using the Alternative Calibration Method.

Please have in mind that we are planning on adding more calibration pipelines to PIX4Dmatic, so that there is more flexibility in regards to calibration in the future.

I have attached the quality report with the processing options we used. If possible I would also set the Internal Parameters Optimization to All Prior .
475_report.pdf (3.3 MB)

I hope you find this helpful!

Best,

Good morning Nikoleta and thanks for reaching back out. Those dataset results look great and what I expected for a forested landscape. Unfortunately, I do not have those processing options in my current trial version of Pix4dMatic to replicate your results. Did you try to process them in Pix4dMatic and what is the expected timeline to add those options to the processing pipeline in 4dMatic? Maybe I should be evaluating Pix4dMapper instead. Than you for your help.

Hi @DH_79, we aim to have an equivalent of the “Alternative” pipeline in PIX4Dmatic in Q1 this year, this is no guarantee but that is what we aim at. I would recommend to have a look at the offers we have online here (if you buy matic, you get a mapper license too): https://www.pix4d.com/pricing/pix4dmatic. That way you should be able to process any kind of project.

Nikoleta is there any way you can share the DTM and Ortho that you generated with me so that I can compare it to the DTM/ortho that I already have?

Good morning Nikoleta, Can you share that DSM and Ortho that you generated with my dataset in Pix4dMapper? Also did you try to process that dataset in Pix4DMatic before finally processing it in Mapper. Let me k kw on both accounts and have a great day! David

Hi,

I am uploading right now the results. I will send you the link once the upload has finished on the personal email you send to us.

Best,

Hi @DH_79

We’ve added new calibration templates and pipelines in the preview 1.25.0 of PIX4Dmatic. It may be worth to try out the following combination for your dataset: Calibration template: “Map” with the Calibration pipeline “Low texture planar”. Although I don’t expect the results to be perfect, it should help to complete the orthomosaic, so that at least you have a full visual overview of the area. Let me know how it went if you try it.

Thanks!

Thank you for the info Pierangelo. Unfortunately, my trial for Pix4DMatic ran out several weeks ago. Any way I could get an extension of the trail so that I can try your suggestions? Thank you again and have a great day!

I’m writing you in PM. A great day to you too!