ViP-NeRF icon indicating copy to clipboard operation
ViP-NeRF copied to clipboard

Running self-made dataset requires DepthBounds.csv, but no instructions on how to obtain this file.

Open rdzhao opened this issue 2 years ago • 1 comments
trafficstars

python VisibilityMask02_NeRF_LLFF.py Program started at 03/11/2023 10:43:47 PM 0%| | 0/1 [00:00<?, ?it/s] ../../../data/databases/SELF/data/all/database_data/scene0001/DepthBounds.csv not found. Traceback (most recent call last): File "/mnt/task_runtime/ViP-NeRF/src/prior_generators/visibility/VisibilityMask02_NeRF_LLFF.py", line 336, in main() File "/mnt/task_runtime/ViP-NeRF/src/prior_generators/visibility/VisibilityMask02_NeRF_LLFF.py", line 328, in main demo1() File "/mnt/task_runtime/ViP-NeRF/src/prior_generators/visibility/VisibilityMask02_NeRF_LLFF.py", line 323, in demo1 start_generation(gen_configs) File "/mnt/task_runtime/ViP-NeRF/src/prior_generators/visibility/VisibilityMask02_NeRF_LLFF.py", line 269, in start_generation bounds = numpy.loadtxt(bounds_path.as_posix(), delimiter=',')[frame_nums] File "/miniconda/envs/vipnerf/lib/python3.9/site-packages/numpy/lib/npyio.py", line 1373, in loadtxt arr = _read(fname, dtype=dtype, comment=comment, delimiter=delimiter, File "/miniconda/envs/vipnerf/lib/python3.9/site-packages/numpy/lib/npyio.py", line 992, in _read fh = np.lib._datasource.open(fname, 'rt', encoding=encoding) File "/miniconda/envs/vipnerf/lib/python3.9/site-packages/numpy/lib/_datasource.py", line 193, in open return ds.open(path, mode, encoding=encoding, newline=newline) File "/miniconda/envs/vipnerf/lib/python3.9/site-packages/numpy/lib/_datasource.py", line 533, in open raise FileNotFoundError(f"{path} not found.") FileNotFoundError: ../../../data/databases/SELF/data/all/database_data/scene0001/DepthBounds.csv not found. Program ended at 03/11/2023 10:43:47 PM Execution time: 0:00:00.385838

FileNotFoundError: ../../../data/databases/SELF/data/all/database_data/scene0001/DepthBounds.csv not found.

I've changed to used my own dataset with some images and CameraIntrinsics.csv and CameraExtrinsics.csv.

rdzhao avatar Nov 03 '23 22:11 rdzhao

DepthBounds.csv contains near and far depth for every image in the scene. It is a csv file with 2 columns (near and far) and number of rows equal to number of images.

Typically, colmap is run to obtain the camera poses of the images. DepthBounds.csv can be extracted from colmap output as done in NerfLlffDataExtractor. This is similar to how NeRF obtains bounds for the scenes.

If the min and max depth are the same across the database, it can be hard-coded as done for VisibilityMask01_RealEstate.py.

NagabhushanSN95 avatar Nov 04 '23 03:11 NagabhushanSN95