How to get camera extrinsic of the BlockNeRF dataset? about unboundednerfpytorch

You can try a simple way as followed, I just tried normalizing the rays_dir in the camera coordinate system and then based on the normalized rays_dir offered by the waymo datasets, we can establish the equation of rays_norm_dir_world=R*rays_norm_dir_cam, this is a superdeterminant equation, and just solve it, after that u can get the proper camera poses. It is worth noting that, despite the simplicity of this method, the average error obtained from the test can reach 1e-5~1e-6

[zymeng1994]

Thanks!

If u still can't figure it, u can email me at [email protected] that I can share u with my script

[zymeng1994]

Thank you very much. I have emailed you.

[sjtuytc]

The detailed scripts with visualizations will be released soon. Thanks and welcome to add my Wechat: ytc407 for coorperation.

[sjtuytc]

@all-contributors please add @SEUleaderYang for providing the camera pose estimation scripts

[allcontributors]

@sjtuytc

I couldn't determine any contributions to add, did you specify any contributions?
Please make sure to use valid contribution names.

[sjtuytc]

@all-contributors please add @SEUleaderYang for code

[allcontributors]

@sjtuytc

I've put up a pull request to add @SEUleaderYang! 🎉

[sjtuytc]

@all-contributors please add @sjtuytc for code

[allcontributors]

@sjtuytc

I've put up a pull request to add @sjtuytc! 🎉

[sjtuytc]

https://github.com/all-contributors please add @sjtuytc for maintenance

[sjtuytc]

@all-contributors please add @sjtuytc for maintenance

[allcontributors]

@sjtuytc

I've put up a pull request to add @sjtuytc! 🎉