总目录
1. 环境及版本说明
系统: UBUNTU20.04
QT: 5.12.12
VTK: 8.2
pcl: 1.11.1
cuda: 11.3
cudnn: 8.2
tensorrt: 8.0.1.6
protobuf 3.20.1
2. 下载并安装Qt
sudo apt-get install mpi-default-dev libicu-dev python-dev libbz2-dev libboost-dev libboost-dev libboost-all-dev gcc g++ pkg-config clang-format-10 libeigen3-dev libflann-dev libusb-dev libusb-1.0 libqhull-dev libglew-dev cmake-gui libflann-dev flex bison libxt-dev libqt5x11extras5-dev
Qt下载链接https://download.qt.io/archive/qt/5.12/5.12.12/qt-opensource-linux-x64-5.12.12.run
下载qt-opensource-linux-x64-5.12.12.run文件,打开命令行,cd到文件所在路径,执行./qt-opensource-linux-x64-5.12.12.run
需要注册qt账号,输入帐号密码,点击next
勾选并点击next
点击next
选择安装路径,点击next
勾选Desktop gcc 64-bit,其他组件根据自己需要进行选择,然后点击next
勾选,点击next
点击install开始安装
安装结束
配置Qt环境变量
sudo gedit ~/.bashrc
# 在末尾加入以下代码,记得把QTDIR更换为自己安装qt的路径
export QTDIR=/usr/local/Qt-5.12.12
export PATH=$QTDIR/bin:$PATH
export MANPATH=$QTDIR/man:$MANPATH
export LD_LIBRARY_PATH=$QTDIR/lib:$LD_LIBRARY_PATH
3. 下载并编译vtk
cd到vtk下载目录
tar xzf VTK-8.2.0.tar.gz
cd VTK-8.2.0
mkdir build
cd build
cmake-gui
选择源码路径和build路径,更改CMAKE_BUILD_TYPE为Release
搜索框输入qt,勾选对应选项,qt相关路径改成你安装qt的路径
点击configure,点击generate,显示Configuring done 和Generating done即可
关闭cmake窗口,控制台输入
make -j8
等待编译完成
sudo make install -j8
4. 下载并编译PCL依赖库
4.1 metslib
metslib下载链接: https://www.coin-or.org/download/source/metslib/metslib-0.5.3.tgz
cd到下载路径
tar xzvf metslib-0.5.3.tgz
cd metslib-0.5.3
./configure
make -j8
sudo make install
4.2 libpng
libpng下载链接: https://sourceforge.net/projects/libpng/files/libpng16/1.6.37/
cd到下载路径
tar xvf libpng-1.6.37.tar.xz
cd libpng-1.6.37
./configure --prefix=/usr/local/libpng
make
sudo make install
4.3 libpcap
libpcap下载链接:http://www.tcpdump.org/#latest-release
cd到下载路径
tar -zxvf libpcap-1.10.1.tar.gz
cd libpcap-1.10.1/
sudo ./configure
sudo make
sudo make install
5. 编译PCL
pcl下载链接:https://github.com/PointCloudLibrary/pcl/releases/download/pcl-1.11.1/source.tar.gz
cd到下载路径
tar -zxvf source.tar.gz
cd pcl
mkdir build
cd build
# 注:如果系统中装了多个qt版本,cmake有可能会找错路径,需要cmake-gui打开手动配置qt路径,
# 然后点击configure generate,然后直接make和sudo make install
cmake -DCMAKE_BUILD_TYPE=Release ..
make -j8
sudo make install -j8
6. cuda+cudnn+tensorrt安装与测试
6.1 cuda11.3安装
下载并安装cuda11.3
wget https://developer.download.nvidia.com/compute/cuda/11.3.0/local_installers/cuda_11.3.0_465.19.01_linux.run
sudo sh cuda_11.3.0_465.19.01_linux.run
配置cuda环境变量
sudo gedit ~/.bashrc
末尾添加如下代码
export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
6.2 cudnn安装
cudnn下载链接: https://developer.nvidia.com/rdp/cudnn-archive
选择如下图所示三个deb文件
依次安装三个包
sudo dpkg -i libcudnn8_8.2.1.32-1+cuda11.3_amd64.deb
sudo dpkg -i libcudnn8-dev_8.2.1.32-1+cuda11.3_amd64.deb
sudo dpkg -i libcudnn8-samples_8.2.1.32-1+cuda11.3_amd64.deb
6.3 TensorRT
tar zxvf TensorRT-8.0.1.6.Linux.x86_64-gnu.cuda-11.3.cudnn8.2.tar.gz
cd TensorRT-8.0.1.6
sudo cp ./lib/* /usr/lib/
sudo cp ./include/* /usr/include/
配置环境变量
sudo gedit ~/.bashrc
末尾添加如下代码,将TensorRT路径换成你自己的TensorRT路径
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/zane/Documents/TensorRT-8.0.1.6/lib
export PATH=$PATH:/home/zane/Documents/TensorRT-8.0.1.6/bin
7. libtorch下载与配置
libtorch下载链接:https://download.pytorch.org/libtorch/lts/1.8/cu111/libtorch-cxx11-abi-shared-with-deps-1.8.2%2Bcu111.zip
如果自己在官网下载记得选择cxx11abi版本
cd到libtorch下载文件夹,解压即可用
unzip libtorch-cxx11-abi-shared-with-deps-1.8.2+cu111.zip
注: 在pcl与libtorch(注意cxx11abi问题)同时使用时,会报libgomp-75eea7e8.so.1: version `GOMP_4.5’ 版本問題,
將系統路徑下的 /usr/lib/x86_64-linux-gnu/libgomp.so.1.0.0 拷貝到 libtorch/lib 路徑下替換 libgomp-75eea7e8.so.1 文件。
8. protobuf
TensorRT依赖protobuf,如果本机没装或者版本较低,需要卸载旧版本重新安装编译
protobuf下载链接: https://github.com/protocolbuffers/protobuf/releases/download/v3.20.1/protobuf-all-3.20.1.tar.gz
cd到protobuf下载路径,执行以下代码
tar zxvf protobuf-all-3.20.1.tar.gz
cd protobuf-3.20.1
./configure
make
sudo make install
9. OpenCV
直接用apt安装
sudo apt install libopencv-dev