Please refer to get_started.md for installation and dataset_prepare.md for dataset preparation.
or like this
my environment: torch=1.9.1,CUDA=11.4
python setup.py build
python setup.py install
pip install mmcv-full==1.5.0 -f https://download.openmmlab.com/mmcv/dist/cu111/torch1.9.0/index.html
- depth image (nyu or kitti)
attention:The depth map names should match Monocular-Depth-Estimation-Toolbox/splits/*.txt
Nyu 3D reconstruction
python ./tools/misc/visualize_point-cloud.py --output-dir point_cloud --dataset nyu --exp_name bts --depth_raw_path "/mnt/data2/datasets/results/nyu/raw/"
Kitti 3D reconstruction
python ./tools/misc/visualize_point-cloud.py --output-dir point_cloud --dataset kitti --exp_name bts --depth_raw_path "/mnt/data2/datasets/results/kitti/raw/"
https://github.com/zhyever/Monocular-Depth-Estimation-Toolbox