Radiometric calibration, DLS correction, and panel correction are performed by preprocess.py.
Project/
├── images
├── panels
└── raw_imagesInput
- raw_images: all images excluding panel
- panel: panel images Output
- images: preprocessed images (radiometric calibration + DLS + panel correction)
source setvars.shWhen QR code can be read panels are detected and read automatically. The program needs to see QR code to locate a panel. If QR code is not visible panel will not be detected and you will need to read it manually. In some cases QR code is visible, but panel albedo QR item may be clipped. It such case program will locate and read panel data, but you will need to provide albedo manually.
(env-micasense) [gnorris@narval1 ImageProcessing]$ python check_panels.py ../Rockland_November_2020/panels/ -fIMG_0773_
...
Error: could not read albedo from 1 panel(s)
Error: 1 panel(s) failed:
Reflectance:
0.5326, 0.5311, 0.5291, 0.5259, 0.5282, 0.5332, 0.5316, nan, 0.5284, 0.5278,
Irradiance:
0.2345, 0.2206, 0.1904, 0.1232, 0.1651, 0.2120, 0.2223, nan, 0.1672, 0.1560, In the example above QR code was clipped and albedo of the panel 8 was missing. In this case you will need to provide albedos of all panels manually.
If there is a message "Panel not found" you will need to read it manually as described in step 3. Otherwise, skip it and proceed to step 4.
To read a panel you need to provide two panel corners and albedo (options -c and -a).
python read_panel.py data/panels_0/IMG_0000_1.tif -c 874 784 704 615 -a 0.53Script needs panel albedo and panel image to calculate panel irradiance:
= * pi / albedo
Radiance is the energy per unit solid angle, W/(sr·m^2). It is determined from images using radiometric calibration parameters stored in metadata.
Irradiance is the radiant flux received by a surface per unit area (W/m^2).
If all panels are OK, script will do it automatically.
If some albedo's can not be read but panels are detected you need to pass a list of albedos in the command line.
python preprocess.py -n773 -d${PROJECT_DIR} -e773 \
-a 0.5326 0.5311 0.5291 0.5259 0.5282 0.5332 0.5316 0.5294 0.5284 0.5278 If some panels are not detected, you need to pass a list of panel irradiances in the command line. You can obtain irradiances of undetected panels with read_panel.py.
module load python/3.10.2 virtualenv env-micasense source env-micasense/bin/activate
LIBZBAR_INSTALL_DIR=$HOME/projects/def-svassili/svassili/ODM/libzbar
git clone https://github.com/mchehab/zbar
cd zbar && autoreconf -vfi
./configure --prefix=${LIBZBAR_INSTALL_DIR} --without-dbus
make install
cd ..
export LD_LIBRARY_PATH=${LIBZBAR_INSTALL_DIR}/lib wget https://exiftool.org/Image-ExifTool-12.67.tar.gz
tar -xf Image-ExifTool-12.67.tar.gz && cd Image-ExifTool-12.67
perl Makefile.PL
make
cd ..
export PATH=$PATH:$HOME/projects/def-svassili/svassili/ODM/Image-ExifTool-12.67- On the Alliance systems python bindings are included in the GDAL module, no installation is required.
module load gcc/9.3.0 opencv/4.8.0 gdal/3.5.1
- On other systems first install GDAL, then install python module:
pip install GDAL==$(gdal-config --version)
git clone https://github.com/micasense/imageprocessing
cd imageprocessing
module load gcc/9.3.0 opencv/4.8.0 gdal/3.5.1
pip install pysolar pyexiftool==0.4.13 pyzbar
pip install --no-index .cd /home/svassili/projects/def-svassili/svassili/ODM/ImageProcessing module load gcc/9.3.0 opencv/4.8.0 gdal/3.5.1 source ../env-micasense/bin/activate export PATH=$PATH:$HOME/projects/def-svassili/svassili/ODM/Image-ExifTool-12.67 export LD_LIBRARY_PATH=$HOME/projects/def-svassili/svassili/ODM/libzbar/lib python Panels.py
'Blue', 'Green', 'Red', 'NIR', 'Red edge'
'Blue-444', 'Green-531', 'Red-650', 'Red edge-705', 'Red edge-740'
increase:
--feature-quality
--min-num-features
--matcher-neighbors
--matcher-distance
--min-num-features 60000
https://community.opendronemap.org/t/hpc-scheduler-e-g-slurm-integration-for-clusterodm/3285/7
apptainer build --fakeroot node_ODM.sif apptainer.def
apptainer overlay create --fakeroot --size 32000 ${SLURM_TMPDIR}/NodeODM.ovl
apptainer run --fakeroot --overlay NodeODM.ovl NodeODM.sif--- File: nodeodm.def ---
Bootstrap: docker
From: opendronemap/odm:latest
%setup
mkdir -p ${APPTAINER_ROOTFS}/var/www
%files
/home/svassili/ODM/NodeODM/* /var/www/
%post
su - root
apt update && apt install -y curl telnet git
curl --silent --location https://deb.nodesource.com/setup_14.x | bash -
apt-get install -y nodejs gdal-bin unzip p7zip-full && npm install -g nodemon && \
ln -s /code/SuperBuild/install/bin/entwine /usr/bin/entwine && \
ln -s /code/SuperBuild/install/bin/entwine /usr/bin/untwine && \
ln -s /code/SuperBuild/install/bin/pdal /usr/bin/pdal
cd /var/www
npm install --production && mkdir -p tmp
%runscript
/bin/bash <<EOF
cd /var/www
node index.jsapptainer build --fakeroot ClusterODM.sif ClusterODM.def
apptainer overlay create --fakeroot --size 3000 ClusterODM.ovl
apptainer run --fakeroot --overlay ClusterODM.ovl ClusterODM.sif --- File: clusterodm.def ---
Bootstrap: docker
From: node:lts
%post
su - root
apt update && apt install -y curl telnet git
git clone https://github.com/OpenDroneMap/ClusterODM
cd ClusterODM
npm install
mv config-default.json config-default.json.org
cat config-default.json.org | sed 's/"admin-pass": "",/"admin-pass": "12345!",/g' > config-default.json
%runscript
/bin/bash <<EOF
cd /ClusterODM
node index.js
EOFtelnet nc11125 8080
LOGIN 12345!
NODE ADD 10.82.90.29 3000
NODE LISTsudo snap install docker
git clone https://github.com/OpenDroneMap/WebODM --config core.autocrlf=input --depth 1
cd WebODM
./webodm.sh start Currently ODM recognized images with up to 8 bands as multispectral.
Types.py can be patched for a 10-band camera:
sed 's/bands_count <= 8:/bands_count <= 10:/g' /code/opendm/types.py- Build sandbox
- Edit types.py
- Build production image from sandbox
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent