线特征---LineMatching代码运行（五）

[1]    https://github.com/dlut-dimt/LineMatching

The code is based on Matlab.

 https://github.com/kailigo/LineSegmentMatching

The code is based on the following two papers. The code is based on OpenCV, and you need CMake to build a project from the code.

@article{li2016hierarchical, title={Hierarchical line matching based on Line--Junction--Line structure descriptor and local homography estimation}, author={Li, Kai and Yao, Jian and Lu, Xiaohu and Li, Li and Zhang, Zhichao}, journal={Neurocomputing}, volume={184}, pages={207--220}, year={2016}, publisher={Elsevier} }

@inproceedings{li2014robust, title={Robust line matching based on ray-point-ray structure descriptor}, author={Li, Kai and Yao, Jian and Lu, Xiaohu}, booktitle={Asian Conference on Computer Vision}, pages={554--569}, year={2014}, organization={Springer} }

[3]    https://github.com/chishuideyu/LBD_and_LineMatching

The code is based on OpenCV ，arpack++.

安装BIAS（没安装完成，由于此库比较老，而且opencv中的函数可代替）

1.BIAS的简单介绍 ---http://www.mip.informatik.uni-kiel.de/tiki-index.php?page=BIAS&highlight=BIAS.

2.How to get it ---http://www.mip.informatik.uni-kiel.de/~wwwadmin/Software/index.html

我下载的是：BIAS-2.8.0.tar.gz

查看imagemagick的版本： convert –version，出现：

Version: ImageMagick 6.7.- -- Q16 http://www.imagemagick.org
Copyright: Copyright (C) - ImageMagick Studio LLC
Features: OpenMP

ImageMagick的源码：   https://github.com/ImageMagick/

下载下来：ImageMagick-master.zip，解压，由于我是在linux安装，查看它的Install-unix.txt，并将其重命名为了：ImageMagick-7.0.

安装过程：

cd ImageMagick-7.0.
./configure
make   //build ImageMagick
sudo make install   //install ImageMagick

make check     //Run tests using the installed ImageMagick（可选指令）

ImageMagick is installs binaries in /../usr/local/bin, libraries in /../usr/local/lib, header files in /../usr/local/include and documentation in /../usr/local/share.

安装ARPACK++

下载：https://github.com/m-reuter/arpackpp

阅读“README.md”，对arpackpp (ARPACK++)有大致的了解，并且主要看其Dependencies依赖项，然后看“INSTALL.md”，进行安装：

安装：

[1]只安装头文件，通过cmake

mkdir build
cd build
cmake ..
sudo make install

-- A library with BLAS API found.
-- Looking for Fortran cheev
-- Looking for Fortran cheev - found
-- A library with LAPACK API found.
-- LAPACK_LIBRARIES: /usr/lib/liblapack.a;/usr/lib/libf77blas.a;/usr/lib/libatlas.a
-- ARPACK_LIB found: /usr/lib/libarpack.a

//--ARPACK 头文件：/usr/local/include/arpackpp/
//将其头文件连接到工程中

[2]System Libraries: GFORTRAN, BLAS, LAPACK, ARPACK，通过ubuntu软件源安装。

$ sudo apt-get install -y gfortran libopenblas-dev liblapack-dev libarpack2-dev

先运行了一下程序，在CMakelists.txt中这句话的执行出现错误

TARGET_LINK_LIBRARIES(LineMatchingLib ${OpenCV_LIBS} arpack /usr/lib/x86_64-linux-gnu/libsuperlu.so arpack++) 

error：

[100%] Building CXX object CMakeFiles/TestLineMatchingAlgorithm.dir/TestLineMatchingAlgorithm.cpp.o
Linking CXX executable TestLineMatchingAlgorithm
libLineMatchingLib.so: undefined reference to `set_default_options'
libLineMatchingLib.so: undefined reference to `dCreate_Dense_Matrix'
libLineMatchingLib.so: undefined reference to `cgstrs'
libLineMatchingLib.so: undefined reference to `sp_ienv'
libLineMatchingLib.so: undefined reference to `sgstrs'
libLineMatchingLib.so: undefined reference to `StatFree'
libLineMatchingLib.so: undefined reference to `debug_'
libLineMatchingLib.so: undefined reference to `Destroy_CompCol_Matrix'
libLineMatchingLib.so: undefined reference to `dCreate_CompCol_Matrix'
libLineMatchingLib.so: undefined reference to `sgstrf'
libLineMatchingLib.so: undefined reference to `dsaupd_'
libLineMatchingLib.so: undefined reference to `dgstrs'
libLineMatchingLib.so: undefined reference to `Destroy_CompCol_Permuted'
libLineMatchingLib.so: undefined reference to `zgstrs'
libLineMatchingLib.so: undefined reference to `cgstrf'
libLineMatchingLib.so: undefined reference to `StatInit'
libLineMatchingLib.so: undefined reference to `zgstrf'
libLineMatchingLib.so: undefined reference to `dcopy_'
libLineMatchingLib.so: undefined reference to `Destroy_SuperNode_Matrix'
libLineMatchingLib.so: undefined reference to `get_perm_c'
libLineMatchingLib.so: undefined reference to `dseupd_'
libLineMatchingLib.so: undefined reference to `Destroy_SuperMatrix_Store'
libLineMatchingLib.so: undefined reference to `dgstrf'
libLineMatchingLib.so: undefined reference to `sp_preorder'
collect2: error: ld returned 1 exit status
make[2]: *** [TestLineMatchingAlgorithm] Error 1
make[1]: *** [CMakeFiles/TestLineMatchingAlgorithm.dir/all] Error 2
make: *** [all] Error 

查看是否安装了superlu：

$ dpkg -L libsuperlu-dev
dpkg-query: package 'libsuperlu-dev' is not installed
Use dpkg --info (= dpkg-deb --info) to examine archive files,
and dpkg --contents (= dpkg-deb --contents) to list their contents.

通过ubuntu软件源安装superlu遇到了困难，出现了错误：

$ sudo apt-get install libsuperlu-dev
Reading package lists... Done
Building dependency tree
Reading state information... Done
E: Unable to locate package libsuperlu-dev

上网搜索后说是

#sudo apt-get update   //更新
    执行完后，问题就解决了。

但是我执行完后，依然没有什么卵用。。。反正，网上说的各种都有，后来想着要不下载下来编译再安装吧：

下载superlu：[SuperLU (Version 5.2.x)]  ---   https://github.com/xiaoyeli/superlu

阅读README，知道有两种方法安装，so我使用CMake安装：

mkdir build
cd build
cmake ..
make
sudo make install

libblas.a
libsuperlu.a
libmatgen.a

//库文件：/usr/local/lib/libsuperlu.a
         /usr/local/lib/pkgconfig/superlu.pc
//头文件：/usr/local/include/supermatrix.h
       /usr/local/include/slu_Cnames.h
/usr/local/include/slu_dcomplex.h
/usr/local/include/slu_scomplex.h
/usr/local/include/slu_util.h
 /usr/local/include/superlu_enum_consts.h
 /usr/local/include/slu_sdefs.h
 /usr/local/include/slu_ddefs.h
/usr/local/include/slu_cdefs.h
 /usr/local/include/slu_zdefs.h

过程中生成了几个测试文件：build/TESTING/s_test.out   # single precision, real
               build/TESTING/d_test.out   # double precision, real
               build/TESTING/c_test.out   # single precision, complex
               build/TESTING/z_test.out   # double precision, complex

由于生成的是静态库libsuperlu.a，不知道怎样连接到自己的库上，感觉用不到就删了，但是库文件都在，如果想看可以参照/home/wj/Downloads/arpackpp/external/SuperLU_5.2.1-build。还得仿照人家写的，最终安了个类似的：

$ sudo apt-get install libsuperlu3-dev

这次终于看见曙光了找到了libsuperlu.so：/usr/lib/libsuperlu.so

$ sudo apt-get install libarpack++-dev

运行结果：