Multiple camera support

[KamiCalcium]

Hi,

Great work!

I am wondering for my own dataset, is it possible to support different intrinsic for each camera?

In instant-ngp, we are able to do it by modify the transforms.json: https://github.com/NVlabs/instant-ngp/discussions/797

In your code I found that by design it is for only one camera, for example, the read_meta function https://github.com/apchenstu/TensoRF/blob/17deeedae5ab4106b30a3295709ec3a8a654c7b1/dataLoader/your_own_data.py#L39 just assume that we only have one kind of intrinsic.

Any suggestions? Thanks a lot!

[apchenstu]

you have to do revised the function in the dataloader because the current codebase doesn't include a similar API to adjust the intrinsic loading

[tallgeeze]

my solution would be modify the dataloader script based on tanksandtemple template. before stack poses, just use the same pattern as pose_fname and img_fname, create such folder structure for all intrinsic(after calibrate distortion. Put the calculation of ray part into the for loop among img_fname, pose_fname in read_meta() function. It works