- Potree is a free open-source WebGL based point cloud renderer for large point clouds.
- For own data, we should first convert it to potree-supported format, and then use Potree for rendering on WebGL.
- This repo introduces how to convert common point cloud data and render it on WebGL.
Environments:
- Ubuntu 18.04/20.04.
- Nodejs
- Potree viewer
- Potreeconvert
Install node.js
Download tar.xz on official website (e.g. nodejs-16.x), and unzip
$ tar xf node-vxx.xx.xx-linux-x64.tar.xz
$ cd node-vxx.xx.xx-linux-x64/
$ ./bin/node -v
vxx.xx.xx
If you want to use node in terminal, 2 methods as follows:
a) Set soft links
ln -s /opt/nodejs/bin/npm /usr/local/bin/
ln -s /opt/nodejs/bin/node /usr/local/bin/
# If not, alternative:
ln -s /opt/nodejs/bin/npm /usr/bin/
ln -s /opt/nodejs/bin/node /usr/bin/
# check if work
$ node --version
$ npm --version
# If remove soft links
rm /usr/local/bin/node
rm /usr/local/bin/npm
# or
rm /usr/bin/node
rm /usr/bin/npm
These method: Sometimes software updates may cause soft links to fail.
b) Set environment variables
# vim/vi to open
vim /etc/profile
# add at the end and save
export NODE_HOME=/opt/nodeJs # use your path
export PATH=$PATH:$NODE_HOME/bin
export NODE_PATH=$NODE_HOME/lib/node_modules
# make it effect and verify
source /etc/profile
node -v
c) Using apt manner
If the above two methods fail, such as donot found command for sudo npm XXX, you can use apt manner.
Ubuntu default installation sources, i.e., sudo apt install nodejs，often only install the old version, with just nodejs and no npm.
Install the latest or specific version (with npm), as follows:
# First, download package ([obtain latest version link](https://github.com/nodesource/distributions/blob/master/README.md))
# Node.js v16.x:
curl -fsSL https://deb.nodesource.com/setup_16.x | sudo -E bash -
# Node.js v14.x:
curl -fsSL https://deb.nodesource.com/setup_14.x | sudo -E bash -
# Second, installation
sudo apt-get install -y nodejs
# Using `node -v` and `npm -v` to check whether the installation is successful.
Here we adopt the apt manner. Other details please refer to [1], [2].
git clone https://github.com/potree/potree.git
# or download stable released version: potree-1.8-release from github
cd potree-1.8
npm install # Install dependencies, as specified in package.json, and create a build in ./build/potree.
Run Potree
cd potree
npm start
Go to http://localhost:1234/ to test the examples.
Each time we should go into potree folder and executive "npm start" to start the potree services.
Thus we can set npm start as a self-starting service. First, create a file: npmstart.sh,
cd ***/path/potree
npm start &
Then save it to the folder: /etc/profile.d/，then reboot.
Factly the viewer link is like http://localhost:1234/examples/choosed_file_name.html.
Other self-starting setting methods can be seen here.
Put self-starting script in /etc/rc.d/rc.local. (/etc/rc.local is the same as /etc/rc.d/rc.local),
e.g. `/mnt/npmstart.sh` at the end of the rc.local file.
Reboot after adding the permission:
chmod +x ***.sh
chmod +x /etc/rc.d/rc.local
For you own point cloud (e.g. .las, .laz, .zst, .bin, etc.), you should convert to potree format by PotreeConverter.
Noted, PotreeConverter only support Ubuntu 20.04.
sudo apt-get update && apt-get install -y gcc g++ gdb cmake python
sudo apt update && apt install -y git
# Lastool
sudo git clone https://github.com/m-schuetz/LAStools.git
&& cd LAStools/LASzip
&& mkdir build && cd build
&& cmake -DCMAKE_BUILD_TYPE=Release ..
&& make
sudo apt-get install libtbb2 libtbb-dev
sudo git clone https://github.com/potree/PotreeConverter.git
&& cd PotreeConverter
&& mkdir build && cd build
&& cmake -DCMAKE_BUILD_TYPE=Release -DLASZIP_INCLUDE_DIRS=/LAStools/LASzip/build/src/liblaszip.so ..
&& make
For Ubuntu 18.04, we use docker to pack the PotreeConverter and call this docker for conversion. This repo gives the Dockerfile of PotreeConverter.
Refere here.
By PotreeConverter, we will obtain the converted potree-format data, i.e., octree.bin, metadata.json and hierarchy.bin. Then, modify one of the examples (html files in Potree/examples/) with the your own path of metadata.json file, see Potreeconverter. Open this html by potree viewer to visulize the point cloud, such as http://localhost:1234/examples/test.html.
To generate it automatically, we build a python file (genhtml.py) to automaticly generate the html files, and a shell script (run.sh) to run the PotreeConverter and genhtml to produce the converted files and html files in docker.
Files:
run.sh: input point cloud data path, output results in potree-1.8/pointclouds
./potreeconvert/Potreeconverter input -o outputDir
genhtml.py: generate html into potree/dataview
Usage:
For example, the .las for view is at: /home/pcloud/potree-1.8/data/test.las
- Build the docker:
sudo docker build -t potreetransform .; - Run the docker:
sudo docker run -it -v /home/pcloud/potree-1.8/data/:/usr/src/app/data/ -v /home/pcloud/potree-1.8/dataview/:/usr/src/app/dataview/ potreetransform /usr/src/app/run.sh test.las
file organization:
Potree:
---- potree-1.8
---- data
---- dataview
---- potreetransform
---- PotreeConverter
---- genhtml.py
---- run.sh
---- Nodejs
In Windows, after install PotreeDesktop, potreeconvert may does not work, probably because of the lack of some dlls. Just download these dlls.