search

marcusvaltonen / python-droneposelib

Python library for Visual-Inertial Odometry (VIO)
MIT License
10 stars 1 forks source link

how to use this Lib? #6

Closed SigmaLGX closed 3 years ago

SigmaLGX commented 3 years ago

Thank you for sharing, but how does this library work

Does it have documentation?

marcusvaltonen commented 3 years ago

Hello, the library is under development, and there is no documentation so far. It will be available in the future.

Right now we are incorporating the solvers in a RANSAC framework. When this is done we will improve the documentation.

To use the minimal solvers:

`import numpy as np import droneposelib as dpl output = dpl.get_valtonenornhag_arxiv_2021_frEfr(np.asfortranarray(x1), np.asfortranarray(x2), np.asfortranarray(R1), np.asfortranarray(R2), use_fast_solver)

output = dpl.get_valtonenornhag_arxiv_2021_fEf(np.asfortranarray(x1), np.asfortranarray(x2), np.asfortranarray(R1), np.asfortranarray(R2))

output = dpl.get_valtonenornhag_arxiv_2021_rEr(np.asfortranarray(x1), np.asfortranarray(x2), np.asfortranarray(R1), np.asfortranarray(R2))`

where x1, x2 are the point correspondences (np.ndarray of correct size), R1, R2 np.ndarray (3x3) are the rotation matrices. For the frEfr solver you also have a boolean option for use_fast_solver -- it is faster, but not as accurate.

SigmaLGX commented 3 years ago

thank you