You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am currently working on a project involving Gaussian Splatting (GS) with data reconstructed from a 360 camera using OpenSfM. However, I've encountered a challenge: GS primarily supports the COLMAP data format, but not the OpenSfM format.
To address this, I am considering converting the OpenSfM 360 camera data into the COLMAP format. My approach involves splitting the spherical view into a cube map format and then loading it into GS.
In relation to this, I came across a relevant issue on OpenSfM's GitHub: #794
It seems that in OpenSfM, the process is executed after creating a perspective view from the 360 camera data. Is this understanding correct?
Additionally, I retrieved the following point coordinate from the SfM feature data:
array([4.9925733e-01, 2.3722187e-01, 1.8578832e-01, 3.5989554e+02], dtype=float32),
However, I am uncertain if this coordinate represents the image below or if it has already been converted to a perspective view.
I would appreciate any insights or suggestions on how to properly convert the data format or verify the coordinates.
Thank you for your time and assistance.
The text was updated successfully, but these errors were encountered:
inuex35
changed the title
feature point coordinate
360 camera feature point coordinate
Dec 25, 2023
Hello,
I am currently working on a project involving Gaussian Splatting (GS) with data reconstructed from a 360 camera using OpenSfM. However, I've encountered a challenge: GS primarily supports the COLMAP data format, but not the OpenSfM format.
To address this, I am considering converting the OpenSfM 360 camera data into the COLMAP format. My approach involves splitting the spherical view into a cube map format and then loading it into GS.
In relation to this, I came across a relevant issue on OpenSfM's GitHub:
#794
It seems that in OpenSfM, the process is executed after creating a perspective view from the 360 camera data. Is this understanding correct?
Additionally, I retrieved the following point coordinate from the SfM feature data:
array([4.9925733e-01, 2.3722187e-01, 1.8578832e-01, 3.5989554e+02], dtype=float32),
However, I am uncertain if this coordinate represents the image below or if it has already been converted to a perspective view.
I would appreciate any insights or suggestions on how to properly convert the data format or verify the coordinates.
Thank you for your time and assistance.
The text was updated successfully, but these errors were encountered: