Creating unwrapped polyhedral maps from equirectangular images. Initial python code for Isocahedral from rdbch
Insipration taken from here. If you find this useful, please consider giving it a star ⭐
This code was developed using python 3.8, however it should run on anything that has >= python 3.6.
To install the requirements, one can simply run:
$ pip install -r requiremets.py2 base classes available:
IcosahedralSamplerclass -- for generating 20 triangular faced polygonDodecahedralSamplerclass -- for generating 12 regular pentagon faced polygon
Syntax for both classes should be similar
Sample usage:
from poly_sampler import IcosahedralSampler
eq_image = imread('./assets/0.png')
sampler = IcosahedralSampler(resolution = 600)
# generate unwrapped maps (as presented above)
unwrapped_image = sampler.unwrap(eq_image, face_offset=0) # change face_offset to rotation_offset for DodecahedralSampler
# generate unwrapped maps and shift they faces are put into the final panorama
unwrapped_image = sampler.unwrap(eq_image, face_offset=2) # change face_offset to rotation_offset for DodecahedralSampler
# create the image of the triangular face
face_image = sampler.get_face_image(face_no=0, eq_image=eq_image)
# sample face colors from an eq image
face_colors = sampler.get_face_rgb(face_no=0, eq_image=eq_image)One can run the provided sample notebook (isocahedron and dodecahedron) to see exactly how the code works.
One can create shift the output image by specifying the face_offset when calling sampler.unwrap() or rotation_offset when calling sampler.unwrap():
This repository also contains a command line utility program that can convert an equirectangular image to an incosahedral projection map:
$ python unwrap.py --input=<path to input> \
--output=<path to output> \
--face_resolution=600 \
--face_offset=0