首先下载源码
根据需要下载下载下面的依赖
该源码通过cmake编译，如果没有可以通过conda下载
conda install cmake
1，安装boost
conda install --channel https://conda.anaconda.org/menpo boost
2，安装libzmq
conda install --channel https://conda.anaconda.org/anaconda pyzmq（发现好像用不了，改源码安装）
源码安装：
git clone git://github.com/zeromq/libzmq.git
cd libzmq./autogen.sh./configure --prefix=/home/sk39/soft/zmqmakemake checkmake install
安装好后将/home/sk39/soft/zmq/lib下的文件copy到/home/sk39/anaconda3/lib下面
3、安装cppzmq
conda install --channel https://conda.anaconda.org/QuantStack cppzmq
4、安装opencv
conda install --channel https://conda.anaconda.org/menpo opencv3
5、安装hdf5
由于conda里面已经有了hdf5
conda install --channel https://conda.anaconda.org/anaconda hdf5
进入rmpe_dataset_server文件夹
mkdir build
cd build
cmake ..
make
如果此时出错
可能会报下面的错误：找不到libhdf5_cpp.so和libicui18n.so.58
/home/sk39/anaconda3/lib/libicui18n.so.58: undefined reference to `__cxa_throw_bad_array_new_length@CXXABI_1.3.8'
/home/sk39/anaconda3/lib/libhdf5_cpp.so: undefined reference to `operator delete(void*, unsigned long)@CXXABI_1.3.9'
collect2: error: ld returned 1 exit status
make[2]: *** [rmpe_dataset_server] Error 1
make[1]: *** [CMakeFiles/rmpe_dataset_server.dir/all] Error 2
make: *** [all] Error 2
解决办法：
1，安装hdf5-cxx
conda install --channel https://conda.anaconda.org/inria-pro-sed hdf5-cxx
2，缺ibicui18n.so.58库，但是我在conda下的lib文件下发现有这个so文件，解决办法：
https://anaconda.org/conda-forge/icu/files下载后copy到conda的lib下。