Hello, I have a problem with creating 3DGS using input from reality capture. I generated a CSV and used the correct method according to the instructions in the documentation (Internal/External camera parameters). I also generated and use .ply pointcloud. I aligned dataset twice in RC and no matter what I do I'm still having this problem. It seems that the camera positions are not taken into account during processing and it brings them all to one point. Or not even so much as positions because the angles of tilt looks like they are taken into account but their location in space is not. Have you guys encountered this problem? Is there something missing? Something I should add? I also tried to change export transformation settings for coordinate system and space in both csv and ply but result was the same. I used such commands to go through the process:
PROCESSING:
ns-process-data realitycapture --data "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki\images" --csv "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki\Glinianki_Wloscianski.csv" --ply "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki\Glinianki_Wloscianski.ply" --output-dir "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki" --max_dataset_size -1
TRAINING:
ns-train splatfacto --data "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki"
Hello, I have a problem with creating 3DGS using input from reality capture. I generated a CSV and used the correct method according to the instructions in the documentation (Internal/External camera parameters). I also generated and use .ply pointcloud. I aligned dataset twice in RC and no matter what I do I'm still having this problem. It seems that the camera positions are not taken into account during processing and it brings them all to one point. Or not even so much as positions because the angles of tilt looks like they are taken into account but their location in space is not. Have you guys encountered this problem? Is there something missing? Something I should add? I also tried to change export transformation settings for coordinate system and space in both csv and ply but result was the same. I used such commands to go through the process: PROCESSING:
ns-process-data realitycapture --data "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki\images" --csv "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki\Glinianki_Wloscianski.csv" --ply "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki\Glinianki_Wloscianski.ply" --output-dir "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki" --max_dataset_size -1
TRAINING:ns-train splatfacto --data "D:\Python Projects\nerfstudio\data\REALITY_CAPTURE\glinianki"