Best config.yaml options for spherical images

[Alotaibiuop]

trafficstars

Thank you for this great tool, I have an issue and some questions if you can help

my issue is that some time only ~195 images added to reconstruction file of 500

questions: whats is the best parameter that i can set in config.yaml for spherical images taken by following camera { "v2 nctech ltd istar pulsar 11000 5500 equirectangular 0.0": { "projection_type": "spherical", "width": 11000, "height": 5500 } }

my projects avg images count ~10K, I split them to (500)s then process them parallel to reduce pipeline time

how I can only align images only with correct rotation and transformation without run the whole pipeline.

[YanNoun]

Hi @Alotaibiuop,

Yes, "projection_type": "spherical" is the way to go for spherical images.

Regarding split and merge, you can use the large tools as in the doc : https://opensfm.org/docs/large.html It will basically split the dataset into smaller dataset, run incremental SfM on them, and then re-align them.

Let me know if that helps,

Yann

[Alotaibiuop]

Hi @YanNoun , Thank you for the response, I was wondering about other parameters in the config.yaml for "spherical" images, what is the best practice.

beside that, I am interested in getting "rotation" field only that is calculated after reconstruction.json (for each shot) is there is a way to get that value without running the whole SfM pipeline

thank you for mentioning large tools, that will help.

Tariq

[YanNoun]

Hi @Alotaibiuop ,

The only specifics we have for spherical is the number of features (16K by default) and the downsampling resolution before feature extraction (4096 by default).

Otherwise, the rest is the same.

Yann