360MonoDepth

Based on 360MonoDepth by Manuel Rey-Area, Mingze Yuan and Christian Richardt

Setup

Tested with Python >= 3.8

1. Clone the repository

git clone git@github.com:patriciogonzalezvivo/360monodepth.git
cd 360monodepth

2. Install C++ dependencies for:

• Ceres 2.0.0
• Eigen 3.3.9
• Glog 0.5.0
• Gflags 2.2.2
• GTest 1.10.0
• OpenCV 4.2.0
• Boost 1.75.0
• pybind11 2.8.1

On linux:

sudo apt install libceres-dev libeigen3-dev libgoogle-glog-dev libgflags-dev libgtest-dev libopencv-dev libboost-all-dev python3-pybind11 pybind11-dev

3. We need to create a conda environment with python 3.8 and build the C++ targets

conda create -f environment.yml
conda activate 360monodepth

4. Build Python bindings

cd cpp
mkdir build && cd build
cmake -DCMAKE_BUILD_TYPE=Release ../
make -j8

5. Copy all the dependent DLL/so files to cpp/python/[dll,so]. For example, in Linux cpp/python/so should contain the following dynamic libraries: libamd.so.2, libcholmod.so.3, libglog.so, libm.so.6, libsuitesparseconfig.so.5, libblas.so.3, libcolamd.so.2, libglog.so.0, libopencv_core.so.4.2, libtbb.so.2, libcamd.so.2, libcxsparse.so.3, libgomp.so.1, libopencv_imgproc.so.4.2, libccolamd.so.2, libgflags.so.2.2, liblapack.so.3, libquadmath.so.0, libceres.so.2, libgfortran.so.5, libmetis.so.5, libspqr.so.2

cd cpp/python
python setup.py build
python setup.py bdist_wheel
pip install dist/instaOmniDepth-0.1.0-cp39-cp39-linux_x86_64.whl 

Running code

# create equirect PNG and npy file usign midas2
python main.py -i data/0001.jpg --persp_monodepth=midas2 --npy
# Create a 3D point cloud from the npy file
python plot.py -r data/0001.jpg -d data/0001_midas2_frustum.npy

# create equirect PNG and npy file usign midas3
python main.py -i data/0001.jpg --persp_monodepth=midas3 --npy
# Create a 3D point cloud from the npy file
python plot.py -r data/0001.jpg -d data/0001_midas3_frustum.npy

# create equirect PNG and npy file usign zoe
python main.py -i data/0001.jpg --persp_monodepth=zoedepth --npy
# Create a 3D point cloud from the npy file
python plot.py -r data/0001.jpg -d data/0001_zoedepth_frustum.npy

Citation

@inproceedings{reyarea2021360monodepth,
	title={{360MonoDepth}: High-Resolution 360{\deg} Monocular Depth Estimation},
	author={Manuel Rey-Area and Mingze Yuan and Christian Richardt},
	booktitle={CVPR},
	year={2022}}