本文详细记录了《视觉SLAM十四讲》书中各章节实践代码在不同项目中的编译问题及解决方案，包括错误原因分析和修复步骤，涉及SLAM、g2o、ceres、opencv等库的版本兼容性问题，以及编译选项设置等。
- slambook-master/ch2
slambook-master/ch2# mkdir build && cd build && cmake .. && make -j8
-- The C compiler identification is GNU 9.4.0
-- The CXX compiler identification is GNU 9.4.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Configuring done
-- Generating done
-- Build files have been written to: /mnt/d/slam14/slambook-master/ch2/build
Scanning dependencies of target helloSLAM
Scanning dependencies of target hello
Scanning dependencies of target hello_shared
[ 37%] Building CXX object CMakeFiles/hello_shared.dir/libHelloSLAM.cpp.o
[ 37%] Building CXX object CMakeFiles/hello.dir/libHelloSLAM.cpp.o
[ 37%] Building CXX object CMakeFiles/helloSLAM.dir/helloSLAM.cpp.o
[ 50%] Linking CXX executable helloSLAM
[ 75%] Linking CXX shared library libhello_shared.so
[ 75%] Linking CXX static library libhello.a
[ 75%] Built target hello
[ 75%] Built target helloSLAM
[ 75%] Built target hello_shared
Scanning dependencies of target useHello
[ 87%] Building CXX object CMakeFiles/useHello.dir/useHello.cpp.o
[100%] Linking CXX executable useHello
[100%] Built target useHello
slambook-master/ch2/build# ./helloSLAM
Hello SLAM!
slambook-master/ch2/build# ./helloSLAM
Hello SLAM!
- slambook-master/ch3
1.编译问题
slambook-master/ch3/visualizeGeometry# mkdir build && cd build && cmake .. && make -j8
……
Scanning dependencies of target visualizeGeometry
[ 50%] Building CXX object CMakeFiles/visualizeGeometry.dir/visualizeGeometry.cpp.o
In file included from /usr/local/include/pangolin/utils/signal_slot.h:3,
from /usr/local/include/pangolin/windowing/window.h:35,
from /usr/local/include/pangolin/display/display.h:34,
from /usr/local/include/pangolin/pangolin.h:38,
from /mnt/d/slam14/slambook-master/ch3/visualizeGeometry/visualizeGeometry.cpp:9:
/usr/local/include/sigslot/signal.hpp:109:79: error: ‘decay_t’ is not a member of ‘std’; did you mean ‘decay’?
109 | constexpr bool is_weak_ptr_compatible_v = detail::is_weak_ptr_compatible<std::decay_t<P>>::value;
| ^~~~~~~
将CMakeLists中set(CMAKE_CXX_FLAGS "-std=c++11")注释掉即可
#set(CMAKE_CXX_FLAGS "-std=c++11")
不过编译后运行./visualizeGeometry会有如下报错：
2.编译问题
slambook-master/ch3/visualizeGeometry/build# ./visualizeGeometry
terminate called after throwing an instance of 'std::runtime_error'
what(): Pangolin X11: Failed to open X display
Aborted
2.解决方案
这个估计是WSL系统没有没有配置界面显示所致，待配置了界面显示估计问题就解决了。
- slambook-master/ch4
1.编译问题
slambook-master/ch4/useSophus/useSophus.cpp:5:10: fatal error: Eigen/Core: No such file or directory
5 | #include <Eigen/Core>
在CMakeLists中添加Eigen库
# 添加Eigen头文件
include_directories( "/usr/include/eigen3" ) #apt命令自动安装的eigen库路径
#include_directories( "/usr/local/include/eigen3" ) #源码编译安装的eigen库路径
2.编译问题
slambook-master/ch4/useSophus/useSophus.cpp编译问题
（1）slambook-master/ch4/useSophus/useSophus.cpp:8:10: fatal error: sophus/so3.h: No such file or directory
8 | #include "sophus/so3.h"
2.解决方案
因新版本Sophus库中头文件名有改变，将代码useSophus.cpp:8中
#include <sophus/so3.h>
#include <sophus/se3.h>
改为
#include <sophus/so3.hpp>
#include <sophus/se3.hpp>
3.编译问题
slambook-master/ch4/useSophus/useSophus.cpp:16:17: error: missing template arguments before ‘SO3_R’
16 | Sophus::SO3 SO3_R(R); // Sophus::SO(3)可以直接从旋转矩阵构造
slambook-master/ch4/useSophus/useSophus.cpp:24:34: error: ‘SO3_v’ was not declared in this scope; did you mean ‘SO3_R’?
24 | cout<<"SO(3) from vector: "<<SO3_v<<endl;
slambook-master/ch4/useSophus/useSophus.cpp:25:38: error: ‘SO3_q’ was not declared in this scope; did you mean ‘SO3_R’?
25 | cout<<"SO(3) from quaternion :"<<SO3_q<<endl;
3.解决方案
Sophus库新版本是模板类，定义时需要指定类型，将代码useSophus.cpp中：
Sophus::SO3 SO3_R(R); // Sophus::SO(3)可以直接从旋转矩阵构造
Sophus::SO3d SO3_v( 0, 0, M_PI/2 ); // 亦可从旋转向量构造
Eigen::Quaterniond q(R); // 或者四元数
Sophus::SO3d SO3_q( q );
改为：
//Sophus::SO3 SO3_R(R); // Sophus::SO(3)可以直接从旋转矩阵构造
Sophus::SO3d SO3_R(R); // Sophus::SO(3)可以直接从旋转矩阵构造
//Sophus::SO3 SO3_v( 0, 0, M_PI/2 ); // 亦可从旋转向量构造
Sophus::SO3d SO3_v( 0, 0, M_PI/2 ); // 亦可从旋转向量构造
Eigen::Quaterniond q(R); // 或者四元数
//Sophus::SO3 SO3_q( q );
Sophus::SO3d SO3_q( q );
4.编译问题
slambook-master/ch4/useSophus/useSophus.cpp:33:33: error: ‘template<class Scalar_, int Options> class Sophus::SO3’ used without template arguments
33 | cout<<"so3 hat=\n"<<Sophus::SO3::hat(so3)<<endl;
4.解决方案
因新版本Sophus库打印需要log()函数：，将代码useSophus.cpp中：
cout<<"SO(3) from matrix: "<<SO3_R<<endl;
改为：
cout<<"SO(3) from matrix: "<<SO3_R.log()<<endl;
/usr/bin/ld: CMakeFiles/useSophus.dir/useSophus.cpp.o: in function `std::make_unsigned<int>::type fmt::v8::detail::to_unsigned<int>(int)':
useSophus.cpp:(.text._ZN3fmt2v86detail11to_unsignedIiEENSt13make_unsignedIT_E4typeES4_[_ZN3fmt2v86detail11to_unsignedIiEENSt13make_unsignedIT_E4typeES4_]+0x23): undefined reference to `fmt::v8::detail::assert_fail(char const*, int, char const*)'
……
collect2: error: ld returned 1 exit status
make[2]: *** [CMakeFiles/useSophus.dir/build.make:97: useSophus] Error 1
make[1]: *** [CMakeFiles/Makefile2:83: CMakeFiles/useSophus.dir/all] Error 2
make: *** [Makefile:91: all] Error 2
5.解决方案
没有链接fmt库，链接上fmt 库就没问题了，将CMakeLists.txt中：
target_link_libraries( useSophus ${Sophus_LIBRARIES} )
改为：
target_link_libraries(useSophus ${Sophus_LIBRARIES} fmt)
- slambook-master/ch5
1.编译问题
/slambook-master/ch5/imageBasics
CMake Error at CMakeLists.txt:8 (find_package):
Could not find a configuration file for package "OpenCV" that is compatible
with requested version "3".
The following configuration files were considered but not accepted:
/usr/local/lib/cmake/opencv4/OpenCVConfig.cmake, version: 4.5.5
-- Configuring incomplete, errors occurred!
See also "/mnt/d/slam14/slambook-master/ch5/imageBasics/build/CMakeFiles/CMakeOutput.log".
1.解决方案
因为按照了opencv4.5.5版本不是opencv3.x.x版本，将CMakeLists.txt中：
find_package( OpenCV 3 REQUIRED )
改为：
find_package( OpenCV REQUIRED )
2.编译问题
slambook-master/ch5/imageBasics/build# ./imageBasics
terminate called after throwing an instance of 'std::logic_error'
what(): basic_string::_M_construct null not valid
Aborted
//TODO有待解决，解决其他问题后一并解决
应该是由于wsl没有安装UI界面，导致编译的程序调用界面显示时不能显示所致。
3.编译问题
CMake Error at CMakeLists.txt:15 (find_package):
By not providing "FindPCL.cmake" in CMAKE_MODULE_PATH this project has
asked CMake to find a package configuration file provided by "PCL", but
CMake did not find one.
Could not find a package configuration file provided by "PCL" with any of
the following names:
PCLConfig.cmake
pcl-config.cmake
Add the installation prefix of "PCL" to CMAKE_PREFIX_PATH or set "PCL_DIR"
to a directory containing one of the above files. If "PCL" provides a
separate development package or SDK, be sure it has been installed.
[ 50%] Building CXX object CMakeFiles/joinMap.dir/joinMap.cpp.o
/mnt/d/slam14/slambook-master/ch5/joinMap/joinMap.cpp:10:10: fatal error: pcl/visualization/pcl_visualizer.h: No such file or directory
10 | #include <pcl/visualization/pcl_visualizer.h>
|
3.解决方案
# eigen
#include_directories( "/usr/include/eigen3/" )
include_directories( "/usr/local/include/eigen3/" )
应该时PCL库没有安装成功，需要重新安装PCL库。
4.编译问题
Scanning dependencies of target joinMap
[ 50%] Building CXX object CMakeFiles/joinMap.dir/joinMap.cpp.o
In file included from /usr/local/include/pcl-1.12/pcl/pcl_macros.h:74,
from /usr/local/include/pcl-1.12/pcl/impl/point_types.hpp:42,
from /usr/local/include/pcl-1.12/pcl/point_types.h:354,
from /mnt/d/slam14/slambook-master/ch5/joinMap/joinMap.cpp:8:
/usr/local/include/pcl-1.12/pcl/pcl_config.h:7:4: error: #error PCL requires C++14 or above
7 | #error PCL requires C++14 or above
4.解决方案
在CMakeLists中添加C++14编译
set( CMAKE_CXX_STANDARD 14)
5.编译问题
slambook-master/ch5/joinMap/build# ./joinMap
请在有pose.txt的目录下运行此程序
5.解决方案
解决办法1：
将make生成的build/joinMap文件复制到pose.txt文件下，在运行./joinMap。
解决办法2：
在joinMap.cpp中：
将ifstream fin("./pose.txt");改为ifstream fin("…/pose.txt");
将boost::format fmt( “./%s/%d.%s” ); 改为boost::format fmt( “…/%s/%d.%s” );
之后保存，重新编译运行即可。
- slambook-master/ch6
1.编译问题
slambook-master/ch6/ceres_curve_fitting/build# make -j16
[ 50%] Building CXX object CMakeFiles/curve_fitting.dir/main.cpp.o
In file included from /usr/local/include/ceres/internal/parameter_dims.h:37,
from /usr/local/include/ceres/internal/autodiff.h:151,
from /usr/local/include/ceres/autodiff_cost_function.h:130,
from /usr/local/include/ceres/ceres.h:37,
from /mnt/d/slam14/slambook-master/ch6/ceres_curve_fitting/main.cpp:3:
/usr/local/include/ceres/internal/integer_sequence_algorithm.h:64:21: error: ‘integer_sequence’ is not a member of ‘std’
64 | struct SumImpl<std::integer_sequence<T, N, Ns...>> {
1.解决方案
在CMakeLists中添加C++14编译
set( CMAKE_CXX_STANDARD 14)
2.编译问题
slambook-master/ch6/g2o_curve_fitting/build
[ 50%] Building CXX object CMakeFiles/curve_fitting.dir/main.cpp.o
In file included from /usr/local/include/g2o/core/base_unary_edge.h:30,
from /mnt/d/slam14/slambook-master/ch6/g2o_curve_fitting/main.cpp:3:
/usr/local/include/g2o/core/base_fixed_sized_edge.h:199:32: error: ‘index_sequence’ is not a member of ‘std’
199 | struct HessianTupleType<std::index_sequence<Ints...>> {
在CMakeLists中添加C++14编译
set( CMAKE_CXX_STANDARD 14)
3.编译问题
/mnt/d/slam14/slambook-master/ch6/g2o_curve_fitting/main.cpp: In function ‘int main(int, char**)’:
/mnt/d/slam14/slambook-master/ch6/g2o_curve_fitting/main.cpp:77:49: error: no matching function for call to ‘g2o::BlockSolver<g2o::BlockSolverTraits<3, 1> >::BlockSolver(g2o::BlockSolver<g2o::BlockSolverTraits<3, 1> >::LinearSolverType*&)’
77 | Block* solver_ptr = new Block( linearSolver ); // 矩阵块求解器
| ^
In file included from /usr/local/include/g2o/core/block_solver.h:206,
from /mnt/d/slam14/slambook-master/ch6/g2o_curve_fitting/main.cpp:4:
/usr/local/include/g2o/core/block_solver.hpp:39:1: note: candidate: ‘g2o::BlockSolver<Traits>::BlockSolver(std::unique_ptr<typename Traits::LinearSolverType>) [with Traits = g2o::BlockSolverTraits<3, 1>; typename Traits::LinearSolverType = g2o::LinearSolver<Eigen::Matrix<double, 3, 3> >]’
39 | BlockSolver<Traits>::BlockSolver(std::unique_ptr<LinearSolverType> linearSolver)
| ^~~~~~~~~~~~~~~~~~~
/usr/local/include/g2o/core/block_solver.hpp:39:68: note: no known conversion for argument 1 from ‘g2o::BlockSolver<g2o::BlockSolverTraits<3, 1> >::LinearSolverType*’ {aka ‘g2o::LinearSolver<Eigen::Matrix<double, 3, 3> >*’} to ‘std::unique_ptr<g2o::LinearSolver<Eigen::Matrix<double, 3, 3> >, std::default_delete<g2o::LinearSolver<Eigen::Matrix<double, 3, 3> > > >’
39 | BlockSolver<Traits>::BlockSolver(std::unique_ptr<LinearSolverType> linearSolver)
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~
/mnt/d/slam14/slambook-master/ch6/g2o_curve_fitting/main.cpp:79:103: error: no matching function for call to ‘g2o::OptimizationAlgorithmLevenberg::OptimizationAlgorithmLevenberg(Block*&)’
79 | g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg( solver_ptr );
| ^
3.解决方案
将源码main.cpp相关部分修改如下：
// 构建图优化，先设定g2o
typedef g2o::BlockSolver< g2o::BlockSolverTraits<3,1> > Block; // 每个误差项优化变量维度为3，误差值维度为1
//Block::LinearSolverType* linearSolver = new g2o::LinearSolverDense<Block::PoseMatrixType>(); // 线性方程求解器
std::unique_ptr<Block::LinearSolverType> linearSolver ( new g2o::LinearSolverDense<Block::PoseMatrixType>());
//Block* solver_ptr = new Block( linearSolver ); // 矩阵块求解器
std::unique_ptr<Block> solver_ptr ( new Block ( std::move(linearSolver))); // 矩阵块求解器
// 梯度下降方法，从GN, LM, DogLeg 中选
//g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg( solver_ptr );
g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( std::move(solver_ptr));
// g2o::OptimizationAlgorithmGaussNewton* solver = new g2o::OptimizationAlgorithmGaussNewton( solver_ptr );
// g2o::OptimizationAlgorithmDogleg* solver = new g2o::OptimizationAlgorithmDogleg( solver_ptr );
g2o::SparseOptimizer optimizer; // 图模型
optimizer.setAlgorithm( solver ); // 设置求解器
optimizer.setVerbose( true ); // 打开调试输出
#参考关于高博十四讲中由于g2o更新出现的问题解决_xiaoshuiyisheng的博客-CSDN博客
- slambook-master/ch7
1.编译问题
CMake Error at CMakeLists.txt:10 (find_package):
Could not find a configuration file for package "OpenCV" that is compatible
with requested version "3.1".
The following configuration files were considered but not accepted:
/usr/local/lib/cmake/opencv4/OpenCVConfig.cmake, version: 4.5.5
1.解决方案
将CMakeLists改为：
#find_package( OpenCV 3.1 REQUIRED )
find_package( OpenCV REQUIRED )
# find_package( OpenCV REQUIRED ) # use this if in OpenCV2
find_package( G2O REQUIRED )
find_package( CSparse REQUIRED )
include_directories(
${OpenCV_INCLUDE_DIRS}
${G2O_INCLUDE_DIRS}
${CSPARSE_INCLUDE_DIR}
#"/usr/include/eigen3/"
"/usr/include/eigen3/"
)
2.编译问题
CSPARSE_LIBRARY
linked by target "pose_estimation_3d3d" in directory /mnt/d/slam14/slambook-master/ch7
linked by target "pose_estimation_3d2d" in directory /mnt/d/slam14/slambook-master/ch7
CMake Warning (dev) in CMakeLists.txt:
Policy CMP0021 is not set: Fatal error on relative paths in
INCLUDE_DIRECTORIES target property. Run "cmake --help-policy CMP0021" for
policy details. Use the cmake_policy command to set the policy and
suppress this warning.
Found relative path while evaluating include directories of
"feature_extraction":
"CSPARSE_INCLUDE_DIR-NOTFOUND"
This warning is for project developers. Use -Wno-dev to suppress it.
CMake Warning (dev) in CMakeLists.txt:
Policy CMP0021 is not set: Fatal error on relative paths in
INCLUDE_DIRECTORIES target property. Run "cmake --help-policy CMP0021" for
policy details. Use the cmake_policy command to set the policy and
suppress this warning.
Found relative path while evaluating include directories of
"pose_estimation_2d2d":
"CSPARSE_INCLUDE_DIR-NOTFOUND"
This warning is for project developers. Use -Wno-dev to suppress it.
CMake Warning (dev) in CMakeLists.txt:
Policy CMP0021 is not set: Fatal error on relative paths in
INCLUDE_DIRECTORIES target property. Run "cmake --help-policy CMP0021" for
policy details. Use the cmake_policy command to set the policy and
suppress this warning.
Found relative path while evaluating include directories of
"triangulation":
"CSPARSE_INCLUDE_DIR-NOTFOUND"
This warning is for project developers. Use -Wno-dev to suppress it.
CMake Warning (dev) in CMakeLists.txt:
Policy CMP0021 is not set: Fatal error on relative paths in
INCLUDE_DIRECTORIES target property. Run "cmake --help-policy CMP0021" for
policy details. Use the cmake_policy command to set the policy and
suppress this warning.
Found relative path while evaluating include directories of
"pose_estimation_3d2d":
"CSPARSE_INCLUDE_DIR-NOTFOUND"
This warning is for project developers. Use -Wno-dev to suppress it.
CMake Warning (dev) in CMakeLists.txt:
Policy CMP0021 is not set: Fatal error on relative paths in
INCLUDE_DIRECTORIES target property. Run "cmake --help-policy CMP0021" for
policy details. Use the cmake_policy command to set the policy and
suppress this warning.
Found relative path while evaluating include directories of
"pose_estimation_3d3d":
"CSPARSE_INCLUDE_DIR-NOTFOUND"
This warning is for project developers. Use -Wno-dev to suppress it.
-- Generating done
CMake Generate step failed. Build files cannot be regenerated correctly.
2.解决方案
未安装CSparse，需要执行以下命令来安装CSparse。
sudo apt-get install libsuitesparse-dev
3.编译问题
[ 10%] Building CXX object CMakeFiles/pose_estimation_3d2d.dir/pose_estimation_3d2d.cpp.o
/mnt/d/slam14/slambook-master/ch7/pose_estimation_3d2d.cpp:12:10: fatal error: g2o/solvers/csparse/linear_solver_csparse.h: No such file or directory
12 | #include <g2o/solvers/csparse/linear_solver_csparse.h>
3.解决方案
g2o安装有问题导致缺少/usr/local/include/g2o/solvers/csparse文件夹而找不到g2o/solvers/csparse/linear_solver_csparse.h头文件,重新安装g2o库即可。
4.编译问题
slambook-master/ch7/feature_extraction.cpp: In function ‘int main(int, char**)’:
/mnt/d/slam14/slambook-master/ch7/feature_extraction.cpp:17:35: error: ‘CV_LOAD_IMAGE_COLOR’ was not declared in this scope
17 | Mat img_1 = imread ( argv[1], CV_LOAD_IMAGE_COLOR );
……
在源码feature_extraction.cpp、pose_estimation_2d2d.cpp、pose_estimation_3d2d.cpp、pose_estimation_3d3d.cpp中添加头文件
#include <opencv2/imgcodecs/legacy/constants_c.h>
#参考链接
5.编译问题
slambook-master/ch7/pose_estimation_2d2d.cpp:155:65: error: ‘CV_FM_8POINT’ was not declared in this scope
155 | fundamental_matrix = findFundamentalMat ( points1, points2, CV_FM_8POINT );
5.解决方案
将源码/slambook-master/ch7/pose_estimation_2d2d.cpp中CV_FM_8POINT改为FM_8POINT即可。如下所示：
//fundamental_matrix = findFundamentalMat ( points1, points2, CV_FM_8POINT );
fundamental_matrix = findFundamentalMat ( points1, points2, FM_8POINT );
#参考链接高翔slam14讲中ch7代码中的一些修改_jywell-CSDN博客
6.编译问题
slambook-master/ch7/pose_estimation_3d2d.cpp:9:
/usr/local/include/g2o/core/base_fixed_sized_edge.h:199:32: error: ‘index_sequence’ is not a member of ‘std’
199 | struct HessianTupleType<std::index_sequence<Ints...>> {
6.解决方案
需要在Cmakelists中更新至c++14，在CMakeLists中添加set( CMAKE_CXX_STANDARD 14)
7.编译问题
slambook-master/ch7/pose_estimation_3d2d.cpp:178:14: error: ‘VertexSBAPointXYZ’ is not a member of ‘g2o’; did you mean ‘VertexPointXYZ’?
178 | g2o::VertexSBAPointXYZ* point = new g2o::VertexSBAPointXYZ();
| ^~~~~~~~~~~~~~~~~
| VertexPointXYZ
7.解决方案
新版g2o库中不再存在原先的VertexSBAPointXYZ，将所有报错的地方改为VertexPointXYZ即可。
8.编译问题
（1）slambook-master/ch7/pose_estimation_3d2d.cpp:152:50: error: no matching function for call to ‘g2o::BlockSolver<g2o::BlockSolverTraits<6, 3> >::BlockSolver(g2o::BlockSolver<g2o::BlockSolverTraits<6, 3> >::LinearSolverType*&)’
152 | Block* solver_ptr = new Block ( linearSolver ); // 矩阵块求解器
（2）slambook-master/ch7/pose_estimation_3d3d.cpp:280:49: error: no matching function for call to ‘g2o::BlockSolver<g2o::BlockSolverTraits<6, 3> >::BlockSolver(g2o::BlockSolver<g2o::BlockSolverTraits<6, 3> >::LinearSolverType*&)’
280 | Block* solver_ptr = new Block( linearSolver ); // 矩阵块求解器
8.解决方案
新版本g2o中API接口变了，有两种修改模式，基本意义相同但第一种更简洁：
（1）slambook-master/ch7/pose_estimation_3d2d.cpp解决方案采用第一种方法：参考增加智能指针 std::unique_ptr，第一种方法是根据新版本中的示例来进行语句修改。如下所示：
// 初始化g2o
typedef g2o::BlockSolver< g2o::BlockSolverTraits<6,3> > Block; // pose 维度为 6, landmark 维度为 3
//Block::LinearSolverType* linearSolver = new g2o::LinearSolverCSparse<Block::PoseMatrixType>(); // 线性方程求解器
std::unique_ptr<Block::LinearSolverType> linearSolver ( new g2o::LinearSolverCSparse<Block::PoseMatrixType>());
//Block* solver_ptr = new Block ( linearSolver ); // 矩阵块求解器
//std::unique_ptr<Block> solver_ptr( new Block( linearSolver) );
std::unique_ptr<Block> solver_ptr ( new Block ( std::move(linearSolver))); // 矩阵块求解器
//g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( solver_ptr );
g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( std::move(solver_ptr));
g2o::SparseOptimizer optimizer;
optimizer.setAlgorithm ( solver );
（2）slambook-master/ch7/pose_estimation_3d3d.cpp解决方案采用第二种方法：根据报错来修改Block的初始化方法，在每个语句对应加入unique_ptr。如下所示：
// 初始化g2o
typedef g2o::BlockSolver< g2o::BlockSolverTraits<6,3> > Block; // pose维度为 6, landmark 维度为 3
Block::LinearSolverType* linearSolver = new g2o::LinearSolverEigen<Block::PoseMatrixType>(); // 线性方程求解器
//Block* solver_ptr = new Block( linearSolver ); // 矩阵块求解器
//std::unique_ptr<Block> solver_ptr ( new Block ( std::move(linearSolver))); // 矩阵块求解器
Block* solver_ptr = new Block( std::unique_ptr<Block::LinearSolverType>(linearSolver) ); // 矩阵块求解器
//g2o::OptimizationAlgorithmGaussNewton* solver = new g2o::OptimizationAlgorithmGaussNewton( solver_ptr );
//g2o::OptimizationAlgorithmGaussNewton* solver = new g2o::OptimizationAlgorithmGaussNewton ( std::move(solver_ptr));
g2o::OptimizationAlgorithmGaussNewton* solver = new g2o::OptimizationAlgorithmGaussNewton( unique_ptr<Block>(solver_ptr) );
g2o::SparseOptimizer optimizer;
optimizer.setAlgorithm( solver );
9.编译问题
/usr/bin/ld: CMakeFiles/pose_estimation_3d2d.dir/pose_estimation_3d2d.cpp.o: undefined reference to symbol '_ZTVN3g2o14VertexPointXYZE'
/usr/bin/ld: /usr/local/lib/libg2o_types_slam3d.so: error adding symbols: DSO missing from command line
collect2: error: ld returned 1 exit status
make[2]: *** [CMakeFiles/pose_estimation_3d2d.dir/build.make:113: pose_estimation_3d2d] Error 1
make[1]: *** [CMakeFiles/Makefile2:169: CMakeFiles/pose_estimation_3d2d.dir/all] Error 2
9.解决方案
（1）DSO missing的原因基本上是因为有库没有连接上，导致未识别。最保险的方法是在最前端将所有库设置为G2O_LIBS：
SET(G2O_LIBS g2o_cli g2o_ext_freeglut_minimal g2o_simulator g2o_solver_slam2d_linear g2o_types_icp g2o_types_slam2d g2o_core g2o_interface g2o_solver_csparse g2o_solver_structure_only g2o_types_sba g2o_types_slam3d g2o_csparse_extension g2o_opengl_helper g2o_solver_dense g2o_stuff g2o_types_sclam2d g2o_parser g2o_solver_pcg g2o_types_data g2o_types_sim3 cxsparse )
（2）由于是在pose_estimation_3d2d中出现问题，所以可以在随后对应的链接中加入${G2O_LIBS}
target_link_libraries( pose_estimation_3d2d
${OpenCV_LIBS}
${CSPARSE_LIBRARY}
g2o_core g2o_stuff g2o_types_sba g2o_csparse_extension ${G2O_LIBS}
)
#参考链接
SLAM十四讲ch7代码调整（undefined reference to symbol）_山馗的博客-CSDN博客
（1）slambook-master/ch7/build# ./feature_extraction
usage: feature_extraction img1 img2
类似的问题还有
slambook-master/ch7/build# ./triangulation
usage: triangulation img1 img2
slambook-master/ch7/build# . /pose_estimation_2d2d
usage: pose_estimation_2d2d img1 img2
slambook-master/ch7/build# . /pose_estimation_3d2d
usage: pose_estimation_3d2d img1 img2 depth1 depth2
slambook-master/ch7/build# . /pose_estimation_3d3d
usage: pose_estimation_3d3d img1 img2 depth1 depth2
（2）slambook-master/ch7/build# ./feature_extraction ../1.png ../2.png
terminate called after throwing an instance of 'cv::Exception'
what(): OpenCV(4.5.5-dev) /mnt/d/slam14/wsl_libs/opencv-4.x/modules/highgui/src/window.cpp:1268: error: (-2:Unspecified error) The function is not implemented. Rebuild the library with Windows, GTK+ 2.x or Cocoa support. If you are on Ubuntu or Debian, install libgtk2.0-dev and pkg-config, then re-run cmake or configure script in function 'cvShowImage'
Aborted
10.解决方案
（1）报错的意思就是以下运行命令格式不对：
./feature_extraction
./triangulation
./pose_estimation_2d2d
./pose_estimation_3d2d
./pose_estimation_3d3d
正确的命令运行格式应该为：
./feature_extraction img1 img2
./triangulation img1 img2
./pose_estimation_2d2d img1 img2
./pose_estimation_3d2d img1 img2 depth1 depth2
./pose_estimation_3d3d img1 img2 depth1 depth2
如果这些二进制文件feature_extraction、triangulation、pose_estimation_2d2d img1 img2、pose_estimation_3d2d、pose_estimation_3d3d是在源代码目录生成的或者是在新建的build目录中编译生成后移动或者拷贝到源码文件目录下的，实际的运行命令为：
./feature_extraction 1.png 2.png
./triangulation 1.png 2.png
./pose_estimation_2d2d 1.png 2.png
./pose_estimation_3d2d 1.png 2.png 1_depth.png 2_depth.png
./pose_estimation_3d3d 1.png 2.png 1_depth.png 2_depth.png
如果这些二进制文件feature_extraction、triangulation、pose_estimation_2d2d img1 img2、pose_estimation_3d2d、pose_estimation_3d3d是在新建的build目录中编译生成的不移动或拷贝它们，实际的运行命令为：
./feature_extraction ../1.png ../2.png
./triangulation ../1.png ../2.png
./pose_estimation_2d2d ../1.png ../2.png
./pose_estimation_3d2d ../1.png ../2.png ../1_depth.png ../2_depth.png
./pose_estimation_3d3d ../1.png ../2.png ../1_depth.png ../2_depth.png
（2）//TODO有待解决，解决其他问题后一并解决
应该是由于wsl没有安装UI界面，导致编译的程序调用界面显示时不能显示所致。
- slambook-master/ch8
1.编译问题
（1）slambook-master/ch8/LKFlow/build# ./useLK
usage: useLK path_to_dataset
（2）slambook-master/ch8/LKFlow/build# ./useLK ../../data/
LK Flow use time：0.0142803 seconds.
tracked keypoints: 1750
terminate called after throwing an instance of 'cv::Exception'
what(): OpenCV(4.5.5-dev) /mnt/d/slam14/wsl_libs/opencv-4.x/modules/highgui/src/window.cpp:1268: error: (-2:Unspecified error) The function is not implemented. Rebuild the library with Windows, GTK+ 2.x or Cocoa support. If you are on Ubuntu or Debian, install libgtk2.0-dev and pkg-config, then re-run cmake or configure script in function 'cvShowImage'
Aborted
1.解决方案
（1）退出LKFlow文件夹，进入data文件夹并将将data文件夹中的压缩文件用tar -zxvf data.tar.gz命令解压，然后进入LKFlow/build文件夹目录执行./useLK ../../data/命令
（2）//TODO有待解决，解决其他问题后一并解决
应该是由于wsl没有安装UI界面，导致编译的程序调用界面显示时不能显示所致。
2.编译问题
[ 50%] Building CXX object CMakeFiles/direct_semidense.dir/direct_semidense.cpp.o
In file included from /usr/local/include/g2o/core/base_unary_edge.h:30,
from /mnt/d/slam14/slambook-master/ch8/directMethod/direct_sparse.cpp:14:
/usr/local/include/g2o/core/base_fixed_sized_edge.h:199:32: error: ‘index_sequence’ is not a member of ‘std’
199 | struct HessianTupleType<std::index_sequence<Ints...>> {
2.解决方案
在CMakeLists中添加C++14编译
set( CMAKE_CXX_STANDARD 14)
3.编译问题
（1）slambook-master/ch8/directMethod/direct_semidense.cpp: In function ‘bool poseEstimationDirect(const std::vector<Measurement>&, cv::Mat*, Eigen::Matrix3f&, Eigen::Isometry3d&)’:
/mnt/d/slam14/slambook-master/ch8/directMethod/direct_semidense.cpp:266:62: error: no matching function for call to ‘g2o::BlockSolver<g2o::BlockSolverTraits<6, 1> >::BlockSolver(g2o::BlockSolver<g2o::BlockSolverTraits<6, 1> >::LinearSolverType*&)’
266 | DirectBlock* solver_ptr = new DirectBlock ( linearSolver );
（2）slambook-master/ch8/directMethod/direct_sparse.cpp:258:62: error: no matching function for call to ‘g2o::BlockSolver<g2o::BlockSolverTraits<6, 1> >::BlockSolver(g2o::BlockSolver<g2o::BlockSolverTraits<6, 1> >::LinearSolverType*&)’
258 | DirectBlock* solver_ptr = new DirectBlock ( linearSolver );
3.解决方案
新版本g2o中API接口变了，有两种修改模式，基本意义相同但第一种更简洁：
（1）slambook-master/ch8/directMethod/direct_semidense.cpp解决方案采用第一种方法：参考增加智能指针 std::unique_ptr，第一种方法是根据新版本中的示例来进行语句修改。如下所示：
// 初始化g2o
typedef g2o::BlockSolver<g2o::BlockSolverTraits<6,1>> DirectBlock; // 求解的向量是6＊1的
//DirectBlock::LinearSolverType* linearSolver = new g2o::LinearSolverDense< DirectBlock::PoseMatrixType > ();
std::unique_ptr<DirectBlock::LinearSolverType> linearSolver ( new g2o::LinearSolverDense<DirectBlock::PoseMatrixType>());
//DirectBlock* solver_ptr = new DirectBlock ( linearSolver );
std::unique_ptr<DirectBlock> solver_ptr ( new DirectBlock ( std::move(linearSolver))); // 矩阵块求解器
// g2o::OptimizationAlgorithmGaussNewton* solver = new g2o::OptimizationAlgorithmGaussNewton( solver_ptr ); // G-N
//g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( solver_ptr ); // L-M
g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( std::move(solver_ptr));
g2o::SparseOptimizer optimizer;
optimizer.setAlgorithm ( solver );
optimizer.setVerbose( true );
（2）slambook-master/ch8/directMethod/direct_sparse.cpp解决方案也采用第一种方法：参考增加智能指针 std::unique_ptr，第一种方法是根据新版本中的示例来进行语句修改。如下所示：
// 初始化g2o
typedef g2o::BlockSolver<g2o::BlockSolverTraits<6,1>> DirectBlock; // 求解的向量是6＊1的
//DirectBlock::LinearSolverType* linearSolver = new g2o::LinearSolverDense< DirectBlock::PoseMatrixType > ();
std::unique_ptr<DirectBlock::LinearSolverType> linearSolver ( new g2o::LinearSolverDense<DirectBlock::PoseMatrixType>());
//DirectBlock* solver_ptr = new DirectBlock ( linearSolver );
std::unique_ptr<DirectBlock> solver_ptr ( new DirectBlock ( std::move(linearSolver))); // 矩阵块求解器
// g2o::OptimizationAlgorithmGaussNewton* solver = new g2o::OptimizationAlgorithmGaussNewton( solver_ptr ); // G-N
//g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( solver_ptr ); // L-M
g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( std::move(solver_ptr));
g2o::SparseOptimizer optimizer;
optimizer.setAlgorithm ( solver );
optimizer.setVerbose( true );
4.编译问题
slambook-master/ch8/directMethod/build# ./direct_semidense
usage: useLK path_to_dataset
4.解决方案
运行命令格式不对，采用useLK path_to_dataset格式如下：
slambook-master/ch8/directMethod/build# ./direct_semidense ../../data/
- slambook-master/ch9
1.编译问题
CMake Error at CMakeLists.txt:16 (find_package):
Could not find a configuration file for package "OpenCV" that is compatible
with requested version "3.1".
The following configuration files were considered but not accepted:
/usr/local/lib/cmake/opencv4/OpenCVConfig.cmake, version: 4.5.5
1.解决方案
将CMakeLists改为：
# OpenCV
#find_package( OpenCV 3.1 REQUIRED )
find_package( OpenCV REQUIRED )
2.编译问题
[ 83%] Building CXX object src/CMakeFiles/myslam.dir/config.cpp.o
In file included from /mnt/d/slam14/slambook-master/project/0.1/src/mappoint.cpp:20:
/mnt/d/slam14/slambook-master/project/0.1/include/myslam/common_include.h:32:10: fatal error: sophus/se3.h: No such file or directory
32 | #include <sophus/se3.h>
2.解决方案
因新版本Sophus库中头文件名有改变，将代码common_include.h:32中
// for Sophus
#include <sophus/se3.hpp>
改为
// for Sophus
//#include <sophus/se3.h>
#include <sophus/se3.hpp>
3.编译问题
In file included from /mnt/d/slam14/slambook-master/project/0.3/src/g2o_types.cpp:1:
/mnt/d/slam14/slambook-master/project/0.3/include/myslam/g2o_types.h:37:80: error: ‘VertexSBAPointXYZ’ is not a member of ‘g2o’; did you mean ‘VertexPointXYZ’?
37 | class EdgeProjectXYZRGBD : public g2o::BaseBinaryEdge<3, Eigen::Vector3d, g2o::VertexSBAPointXYZ, g2o::VertexSE3Expmap>
3.解决方案
将g2o_types.cpp 和g2o_types.h中‘VertexSBAPointXYZ’修改为 ‘VertexPointXYZ’
4.编译问题
/mnt/d/slam14/slambook-master/project/0.3/src/visual_odometry.cpp:202:26: error: ‘class Sophus::SE3<double>’ has no member named ‘rotation_matrix’; did you mean ‘rotationMatrix’?
202 | T_c_r_estimated_.rotation_matrix(),
| ^~~~~~~~~~~~~~~
4.解决方案
将visual_odometry.cpp中‘rotation_matrix’修改为‘rotationMatrix’
5.编译问题
/mnt/d/slam14/slambook-master/project/0.3/src/visual_odometry.cpp:194:49: error: no matching function for call to ‘g2o::BlockSolver<g2o::BlockSolverTraits<6, 2> >::BlockSolver(g2o::BlockSolver<g2o::BlockSolverTraits<6, 2> >::LinearSolverType*&)’
194 | Block* solver_ptr = new Block( linearSolver );
5.解决方案
将visual_odometry.cpp中相关代码修改如下：
// using bundle adjustment to optimize the pose
typedef g2o::BlockSolver<g2o::BlockSolverTraits<6,2>> Block;
//Block::LinearSolverType* linearSolver = new g2o::LinearSolverDense<Block::PoseMatrixType>();
std::unique_ptr<Block::LinearSolverType> linearSolver ( new g2o::LinearSolverDense<Block::PoseMatrixType>());
//Block* solver_ptr = new Block( linearSolver );
std::unique_ptr<Block> solver_ptr ( new Block ( std::move(linearSolver))); // 矩阵块求解器
//g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( solver_ptr );
g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( std::move(solver_ptr));
g2o::SparseOptimizer optimizer;
optimizer.setAlgorithm ( solver );
6.编译问题
（1）[ 50%] Building CXX object src/CMakeFiles/myslam.dir/camera.cpp.o
In file included from /mnt/d/slam14/slambook-master/project/0.1/include/myslam/frame.h:24,
from /mnt/d/slam14/slambook-master/project/0.1/include/myslam/map.h:24,
from /mnt/d/slam14/slambook-master/project/0.1/src/map.cpp:20:
/mnt/d/slam14/slambook-master/project/0.1/include/myslam/camera.h:41:55: error: invalid use of template-name ‘Sophus::SE3’ without an argument list
41 | Vector3d world2camera( const Vector3d& p_w, const SE3& T_c_w );
（2）[ 83%] Building CXX object src/CMakeFiles/myslam.dir/config.cpp.o
In file included from /mnt/d/slam14/slambook-master/project/0.1/src/frame.cpp:20:
/mnt/d/slam14/slambook-master/project/0.1/include/myslam/frame.h:37:5: error: invalid use of template-name ‘Sophus::SE3’ without an argument list
37 | SE3 T_c_w_; // transform from world to camera
（3）[ 50%] Building CXX object src/CMakeFiles/myslam.dir/camera.cpp.o
/mnt/d/slam14/slambook-master/project/0.1/src/camera.cpp:29:60: error: invalid use of template-name ‘Sophus::SE3’ without an argument list
29 | Vector3d Camera::world2camera ( const Vector3d& p_w, const SE3& T_c_w )
（4）[ 33%] Building CXX object src/CMakeFiles/myslam.dir/camera.cpp.o
/mnt/d/slam14/slambook-master/project/0.1/src/frame.cpp:30:44: error: ‘SE3’ is not a type
30 | Frame::Frame ( long id, double time_stamp, SE3 T_c_w, Camera::Ptr camera, Mat color, Mat depth )
| ^~~
/mnt/d/slam14/slambook-master/project/0.1/src/frame.cpp:30:1: error: no declaration matches ‘myslam::Frame::Frame(long int, double, int, myslam::Camera::Ptr, cv::Mat, cv::Mat)’
30 | Frame::Frame ( long id, double time_stamp, SE3 T_c_w, Camera::Ptr camera, Mat color, Mat depth )
（5）/mnt/d/slam14/slambook-master/project/0.2/test/run_vo.cpp:79:9: error: ‘SE3’ was not declared in this scope; did you mean ‘Sophus::SE3’?
79 | SE3 Tcw = pFrame->T_c_w_.inverse();
| ^~~
| Sophus::SE3
In file included from /mnt/d/slam14/slambook-master/project/0.2/include/myslam/common_include.h:34,
from /mnt/d/slam14/slambook-master/project/0.2/include/myslam/config.h:23,
from /mnt/d/slam14/slambook-master/project/0.2/test/run_vo.cpp:8:
/usr/local/include/sophus/se3.hpp:11:7: note: ‘Sophus::SE3’ declared here
11 | class SE3;
| ^~~
（6）/mnt/d/slam14/slambook-master/project/0.2/src/visual_odometry.cpp: In member function ‘void myslam::VisualOdometry::poseEstimationPnP()’:
/mnt/d/slam14/slambook-master/project/0.2/src/visual_odometry.cpp:181:27: error: missing template arguments before ‘(’ token
181 | T_c_r_estimated_ = SE3(
| ^
/mnt/d/slam14/slambook-master/project/0.2/src/visual_odometry.cpp:182:12: error: missing template arguments before ‘(’ token
182 | SO3(rvec.at<double>(0,0), rvec.at<double>(1,0), rvec.at<double>(2,0)),
| ^
6.解决方案
// project中0.1、0.2、0.3、0.4项目中均有此问题。
由于Sophus::SE3和Sophus::SO3在Sophus库新版本中都是模板类，需要声明模板类的数据类型，有两种办法：
（1）一种办法是
using Sophus::SE3;
using Sophus::SO3;
SE3<double> xxx;
SO3<double> yyy;
（2）另一种办法是
using Sophus::SE3d;
using Sophus::SO3d;
SE3d xxx;
SO3d yyy;
采用第二种方法后可将const SE3& T_c_w改为：const SE3d & T_c_w，其他类似地把SE3改为SE3d。最难修改的是T_c_r_estimated_初始化部分，采用第二种方法修改源码如下：
//T_c_r_estimated_ = SE3(
// SO3(rvec.at<double>(0,0), rvec.at<double>(1,0), rvec.at<double>(2,0)),
// Vector3d( tvec.at<double>(0,0), tvec.at<double>(1,0), tvec.at<double>(2,0))
// 此处旋转向量经罗德里格斯转换
Mat R;
cv::Rodrigues(rvec, R);
Eigen::Matrix3d RE;
RE << R.at<double>(0,0), R.at<double>(0,1), R.at<double>(0,2),
R.at<double>(1,0), R.at<double>(1,1), R.at<double>(1,2),
R.at<double>(2,0), R.at<double>(2,1), R.at<double>(2,2);
// SO3构造函数参数为旋转矩阵
T_c_r_estimated_ = SE3d(
SO3d(RE),
Vector3d(tvec.at<double>(0,0), tvec.at<double>(1,0), tvec.at<double>(2,0)));
#参考链接
关于使用新版Sophus(2018年6月github版本)的使用_王勇的博客-CSDN博客
视觉SLAM十四讲 从理论到实践-第九讲实践：设计前端，关于Sophus库中SO3类构造函数使用疑惑_RobotLife的博客-CSDN博客
[ 88%] Building CXX object test/CMakeFiles/run_vo.dir/run_vo.cpp.o
/mnt/d/slam14/slambook-master/project/0.2/test/run_vo.cpp:6:10: fatal error: opencv2/viz.hpp: No such file or directory
6 | #include <opencv2/viz.hpp>
7.解决方案
这是由于编译VTK+opencv+opencv_contrib不完整所致，重新编译opencv+opencv_contrib即可，编译命令中要开VTK：-D WITH_VTK=ON。
8.编译问题
/mnt/d/slam14/slambook-master/project/0.2/test/run_vo.cpp:84:17: error: ‘Tcw’ was not declared in this scope
84 | Tcw.rotation_matrix()(0,0), Tcw.rotation_matrix()(0,1), Tcw.rotation_matrix()(0,2),
/mnt/d/slam14/slambook-master/project/0.2/test/run_vo.cpp:84:21: error: ‘using SE3d = class Sophus::SE3<double>’ {aka ‘class Sophus::SE3<double>’} has no member named ‘
rotation_matrix’; did you mean ‘rotationMatrix’?
84 | Tcw.rotation_matrix()(0,0), Tcw.rotation_matrix()(0,1), Tcw.rotation_matrix()(0,2),
8.解决方案
将‘rotation_matrix’修改为‘rotationMatrix’即可。
9.编译问题
[100%] Linking CXX executable ../../bin/run_vo
/usr/bin/ld: CMakeFiles/run_vo.dir/run_vo.cpp.o: in function `char const* fmt::v8::detail::parse_replacement_field<char, fmt::v8::detail::format_string_checker<char, fmt::v8::detail::error_handler, Eigen::Transpose<Eigen::Matrix<double, 4, 1, 0, 4, 1> > >&>(char const*, char const*, fmt::v8::detail::format_string_checker<char, fmt::v8::detail::error_handler, Eigen::Transpose<Eigen::Matrix<double, 4, 1, 0, 4, 1> > >&)':
run_vo.cpp:(.text._ZN3fmt2v86detail23parse_replacement_fieldIcRNS1_21format_string_checkerIcNS1_13error_handlerEJN5Eigen9TransposeINS5_6MatrixIdLi4ELi1ELi0ELi4ELi1EEEEEEEEEEPKT_SE_SE_OT0_[_ZN3fmt2v86detail23parse_replacement_fieldIcRNS1_21format_string_checkerIcNS1_13error_handlerEJN5Eigen9TransposeINS5_6MatrixIdLi4ELi1ELi0ELi4ELi1EEEEEEEEEEPKT_SE_SE_OT0_]+0x163): undefined reference to `fmt::v8::detail::error_handler::on_error(char const*)'
9.解决方案
没有链接fmt库，链接上fmt 库就没问题了，将CMakeLists.txt中：
add_executable( run_vo run_vo.cpp )
target_link_libraries( run_vo myslam )
改为：
add_executable( run_vo run_vo.cpp )
target_link_libraries( run_vo myslam fmt)
10.编译问题
/mnt/d/slam14/slambook-master/project/0.4/src/g2o_types.cpp:7:16: error: ‘SE3’ in namespace ‘g2o’ does not name a type
7 | const g2o::SE3* point = static_cast<const g2o::SE3*> ( _vertices[0] );
| ^~~
/mnt/d/slam14/slambook-master/project/0.4/src/g2o_types.cpp:9:52: error: ‘point’ was not declared in this scope
9 | _error = _measurement - pose->estimate().map ( point->estimate() );
| ^~~~~
10.解决方案
在头文件g2o_types.h中添加以下头文件引用即可：
#include <g2o/core/base_vertex.h>
#include <g2o/core/base_binary_edge.h>
#include <g2o/core/block_solver.h>
#include <g2o/core/optimization_algorithm_levenberg.h>
#include <g2o/solvers/eigen/linear_solver_eigen.h>
#include <g2o/types/slam3d/types_slam3d.h>
#include <g2o/types/slam3d/types_slam3d.h>
#include <g2o/core/sparse_optimizer.h>
#include <g2o/core/block_solver.h>
#include <g2o/core/factory.h>
#include <g2o/core/optimization_algorithm_factory.h>
#include <g2o/core/optimization_algorithm_gauss_newton.h>
#include <g2o/core/robust_kernel.h>
#include <g2o/core/robust_kernel_factory.h>
#include <g2o/core/optimization_algorithm_levenberg.h>
#include <g2o/solvers/eigen/linear_solver_eigen.h>
#include <g2o/solvers/csparse/linear_solver_csparse.h>
#参考链接：ORB_SLAM之error: ‘Sim3’ in namespace ‘g2o’ does not name a type简单暴力解决问题_英雄小摔哥的博客-CSDN博客
11.编译问题
[100%] Linking CXX executable ../../bin/run_vo
/usr/bin/ld: ../../bin/run_vo: hidden symbol `_ZNK3fmt2v86detail10locale_ref3getISt6localeEET_v' in /usr/local/lib/libfmt.a(format.cc.o) is referenced by DSO
/usr/bin/ld: final link failed: bad value
collect2: error: ld returned 1 exit status
make[2]: *** [test/CMakeFiles/run_vo.dir/build.make:154: ../bin/run_vo] Error 1
make[1]: *** [CMakeFiles/Makefile2:142: test/CMakeFiles/run_vo.dir/all] Error 2
make: *** [Makefile:91: all] Error 2
11.解决方案
本以为安装将gcc，g++版本升级到6.5之后解决了，一查gcc/g++版本是9.4.0的适配Ubuntu20.04LTS版本的，故此法不可行。
#参考链接
后来考虑到跟fmt库兼容问题，重新安装Sophus库（编译命令中添加"-DUSE_BASIC_LOGGING=ON"）编译问题依旧，故此法也行不通。
#参考链接hidden symbol `_ZNK3fmt2v86detail10locale_ref3getISt6localeEET_v‘_一个有梦想，有追求，有抱负的人-CSDN博客
//TODO有待解决，0.2、0.3、0.4项目中均有此编译问题
12.编译问题
[ 20%] Building CXX object src/CMakeFiles/myslam.dir/visual_odometry.cpp.o
In file included from /usr/local/include/g2o/solvers/csparse/csparse_helper.h:30,
from /usr/local/include/g2o/solvers/csparse/linear_solver_csparse.h:32,
from /mnt/d/slam14/slambook-master/project/0.4/include/myslam/g2o_types.h:51,
from /mnt/d/slam14/slambook-master/project/0.4/src/g2o_types.cpp:1:
/usr/local/include/g2o/solvers/csparse/csparse_extension.h:27:10: fatal error: cs.h: No such file or directory
27 | #include <cs.h>
| ^~~~~~
compilation terminated.
make[2]: *** [src/CMakeFiles/myslam.dir/build.make:146: src/CMakeFiles/myslam.dir/g2o_types.cpp.o] Error 1
make[2]: *** Waiting for unfinished jobs....
In file included from /usr/local/include/g2o/solvers/csparse/csparse_helper.h:30,
from /usr/local/include/g2o/solvers/csparse/linear_solver_csparse.h:32,
from /mnt/d/slam14/slambook-master/project/0.4/include/myslam/g2o_types.h:51,
from /mnt/d/slam14/slambook-master/project/0.4/src/visual_odometry.cpp:28:
/usr/local/include/g2o/solvers/csparse/csparse_extension.h:27:10: fatal error: cs.h: No such file or directory
27 | #include <cs.h>
| ^~~~~~
compilation terminated.
make[2]: *** [src/CMakeFiles/myslam.dir/build.make:160: src/CMakeFiles/myslam.dir/visual_odometry.cpp.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:116: src/CMakeFiles/myslam.dir/all] Error 2
make: *** [Makefile:91: all] Error 2
12.解决方案
//TODO有待解决， 0.4项目中均有此编译问题
- slambook-master/ch10
1.编译问题
/mnt/d/slam14/slambook-master/ch10/ceres_custombundle/common/BALProblem.cpp:8:10: fatal error: Eigen/Core: No such file or directory
8 | #include <Eigen/Core>
1.解决方案
在CMakeLists中添加"/usr/include/eigen3"，如下所示：
include_directories(${CERES_INCLUDE_DIRS}
${PROJECT_SOURCE_DIR}/common
${PROJECT_SOURCE_DIR}/common/tools
${PROJECT_SOURCE_DIR}/common/flags
"/usr/include/eigen3" )
2.编译问题
[ 83%] Building CXX object CMakeFiles/ceres_customBundle.dir/ceresBundle.cpp.o
In file included from /usr/local/include/ceres/internal/parameter_dims.h:37,
from /usr/local/include/ceres/internal/autodiff.h:151,
from /usr/local/include/ceres/autodiff_cost_function.h:130,
from /usr/local/include/ceres/ceres.h:37,
from /mnt/d/slam14/slambook-master/ch10/ceres_custombundle/ceresBundle.cpp:3:
/usr/local/include/ceres/internal/integer_sequence_algorithm.h:64:21: error: ‘integer_sequence’ is not a member of ‘std’
64 | struct SumImpl<std::integer_sequence<T, N, Ns...>> {
2.解决方案
在CMakeLists中添加set( CMAKE_CXX_STANDARD 14)
3.编译问题
/mnt/d/slam14/slambook-master/ch10/ceres_custombundle/ceresBundle.cpp:17:14: error: ‘struct ceres::Solver::Options’ has no member named ‘num_linear_solver_threads’; did you mean ‘linear_solver_type’?
17 | options->num_linear_solver_threads = params.num_threads;
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| linear_solver_type
3.解决方案
规避的话注释掉options->num_linear_solver_threads = params.num_threads;
//options->num_linear_solver_threads = params.num_threads;
更好的方法是用“options->num_threads = params.num_threads;”替换原来的“options->num_linear_solver_threads = params.num_threads;”
这是因为“Solver::Options::num_linear_solver_threads is deprecated, Solver::Options::num_threads controls all parallelism in Ceres Solver now.“
4.编译问题
slambook-master/ch10/ceres_custombundle/build# ./ceres_customBundle
Usage: bundle_adjuster -input <path for dataset>
4.解决方案
运行命令应该按照bundle_adjuster -input <path for dataset>，故修改运行命令如下：
./ceres_customBundle -input ../data/problem-16-22106-pre.txt
5.编译问题
[ 83%] Building CXX object CMakeFiles/g2o_customBundle.dir/g2o_bundle.cpp.o
In file included from /usr/local/include/g2o/types/sba/edge_project_psi2uv.h:30,
from /usr/local/include/g2o/types/sba/types_six_dof_expmap.h:38,
from /mnt/d/slam14/slambook-master/ch10/g2o_custombundle/g2o_bundle.cpp:25:
/usr/local/include/g2o/core/base_fixed_sized_edge.h:199:32: error: ‘index_sequence’ is not a member of ‘std’
199 | struct HessianTupleType<std::index_sequence<Ints...>> {
5.解决方案
在CMakeLists中添加set( CMAKE_CXX_STANDARD 14)
[ 83%] Building CXX object CMakeFiles/g2o_customBundle.dir/g2o_bundle.cpp.o
/mnt/d/slam14/slambook-master/ch10/g2o_custombundle/g2o_bundle.cpp: In function ‘void SetMinimizerOptions(std::shared_ptr<g2o::BlockSolver<g2o::BlockSolverTraits<9, 3> > >&, const BundleParams&, g2o::SparseOptimizer*)’:
/mnt/d/slam14/slambook-master/ch10/g2o_custombundle/g2o_bundle.cpp:136:74: error: no matching function for call to ‘g2o::OptimizationAlgorithmLevenberg::OptimizationAlgorithmLevenberg(std::__shared_ptr<g2o::BlockSolver<g2o::BlockSolverTraits<9, 3> >, __gnu_cxx::_S_atomic>::element_type*)’
136 | solver = new g2o::OptimizationAlgorithmLevenberg(solver_ptr.get());
6.解决方案
//this function is unused yet..
//void SetMinimizerOptions(std::shared_ptr<BalBlockSolver>& solver_ptr, const BundleParams& params, g2o::SparseOptimizer* optimizer)
void SetMinimizerOptions(std::unique_ptr<BalBlockSolver>& solver_ptr, const BundleParams& params, g2o::SparseOptimizer* optimizer)
{
//std::cout<<"Set Minimizer .."<< std::endl;
g2o::OptimizationAlgorithmWithHessian* solver;
if(params.trust_region_strategy == "levenberg_marquardt"){
//solver = new g2o::OptimizationAlgorithmLevenberg(solver_ptr.get());
solver = new g2o::OptimizationAlgorithmLevenberg(std::move(solver_ptr));
}
else if(params.trust_region_strategy == "dogleg"){
//solver = new g2o::OptimizationAlgorithmDogleg(solver_ptr.get());
solver = new g2o::OptimizationAlgorithmDogleg(std::move(solver_ptr));
}
else
{
std::cout << "Please check your trust_region_strategy parameter again.."<< std::endl;
exit(EXIT_FAILURE);
}
optimizer->setAlgorithm(solver);
//std::cout<<"Set Minimizer .."<< std::endl;
}
void SetSolverOptionsFromFlags(BALProblem* bal_problem, const BundleParams& params, g2o::SparseOptimizer* optimizer)
{
//BalBlockSolver* solver_ptr;
//g2o::LinearSolver<BalBlockSolver::PoseMatrixType>* linearSolver = 0;
std::unique_ptr<BalBlockSolver::LinearSolverType> linearSolver;
if(params.linear_solver == "dense_schur" ){
//linearSolver = new g2o::LinearSolverDense<BalBlockSolver::PoseMatrixType>();
linearSolver = g2o::make_unique<g2o::LinearSolverDense<BalBlockSolver::PoseMatrixType> >();
}
else if(params.linear_solver == "sparse_schur"){
//linearSolver = new g2o::LinearSolverCholmod<BalBlockSolver::PoseMatrixType>();
//dynamic_cast<g2o::LinearSolverCholmod<BalBlockSolver::PoseMatrixType>* >(linearSolver)->setBlockOrdering(true); // AMD ordering , only needed for sparse cholesky solver
auto cholesky = g2o::make_unique<g2o::LinearSolverCholmod<BalBlockSolver::PoseMatrixType> >();
cholesky->setBlockOrdering(true);
linearSolver = std::move(cholesky);
}
//solver_ptr = new BalBlockSolver(linearSolver);
std::unique_ptr<BalBlockSolver> solver_ptr( new BalBlockSolver(std::move(linearSolver)));
//SetLinearSolver(solver_ptr, params);
//SetMinimizerOptions(solver_ptr, params, optimizer);
g2o::OptimizationAlgorithmWithHessian* solver;
if(params.trust_region_strategy == "levenberg_marquardt"){
//solver = new g2o::OptimizationAlgorithmLevenberg(solver_ptr);
solver = new g2o::OptimizationAlgorithmLevenberg(std::move(solver_ptr));
}
else if(params.trust_region_strategy == "dogleg"){
//solver = new g2o::OptimizationAlgorithmDogleg(solver_ptr);
solver = new g2o::OptimizationAlgorithmDogleg(std::move(solver_ptr));
}
else
{
std::cout << "Please check your trust_region_strategy parameter again.."<< std::endl;
exit(EXIT_FAILURE);
}
optimizer->setAlgorithm(solver);
}
7.编译问题
slambook-master/ch10/g2o_custombundle/ceres/autodiff.h:225:11: error: ‘class ceres::internal::FixedArray<ceres::Jet<double, 12>, 17, Eigen::aligned_allocator<ceres::Jet<double, 12> > >’ has no member named ‘get’
225 | x.get() + jet0,
| ~~^~~
7.解决方案
//TODO有待解决
8.编译问题
slambook-master/ch10/g2o_custombundle/g2o_bundle.cpp:169:63: required from here
/usr/include/c++/9/ext/new_allocator.h:146:4: error: no matching function for call to ‘g2o::BlockSolver<g2o::BlockSolverTraits<9, 3> >::BlockSolver(g2o::LinearSolver<Eigen::Matrix<double, 9, 9, 0, 9, 9> >*&)’
146 | { ::new((void *)__p) _Up(std::forward<_Args>(__args)...); }
8.解决方案
//TODO未解决
- slambook-master/ch11
1.编译问题
（1）slambook-master/ch11/pose_graph_gtsam.cpp:6:10: fatal error: sophus/se3.h: No such file or directory
6 | #include <sophus/se3.h>
| ^~~~~~~~~~~~~~
（2）slambook-master/ch11/pose_graph_g2o_lie_algebra.cpp:15:10: fatal error: sophus/se3.h: No such file or directory
15 | #include <sophus/se3.h>
1.解决方案
（1）将slambook-master/ch11/pose_graph_gtsam.cpp:6中代码修改如下：
//#include <sophus/se3.h>
//#include <sophus/so3.h>
#include <sophus/se3.hpp>
#include <sophus/so3.hpp>
（2）将slambook-master/ch11/pose_graph_g2o_lie_algebra.cpp:15中代码修改如下：
//#include <sophus/se3.h>
//#include <sophus/so3.h>
#include <sophus/se3.hpp>
#include <sophus/so3.hpp>
2.编译问题
[ 66%] Built target pose_graph_gtsam
In file included from /usr/local/include/g2o/core/base_binary_edge.h:30,
from /mnt/d/slam14/slambook-master/ch11/pose_graph_g2o_lie_algebra.cpp:7:
/usr/local/include/g2o/core/base_fixed_sized_edge.h:199:32: error: ‘index_sequence’ is not a member of ‘std’
199 | struct HessianTupleType<std::index_sequence<Ints...>> {
2.解决方案
在CMakeLists中添加C++14编译
set( CMAKE_CXX_STANDARD 14)
3.编译问题
slambook-master/ch11/pose_graph_g2o_lie_algebra.cpp:33:21: error: missing template arguments before ‘e’
33 | Matrix6d JRInv( SE3 e )
3.解决方案
将SE3修改为SE3<double>
类似的问题将SO3修改为SO3<double>
4.编译问题
slambook-master/ch11/pose_graph_g2o_SE3.cpp:38:103: error: no matching function for call to ‘g2o::OptimizationAlgorithmLevenberg::OptimizationAlgorithmLevenberg(Block*&)’
38 | g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg( solver_ptr );
4.解决方案
将pose_graph_g2o_lie_algebra.cpp和pose_graph_g2o_SE3.cpp中相关代码修改为：
typedef g2o::BlockSolver<g2o::BlockSolverTraits<6,6>> Block; // 6x6 BlockSolver
//Block::LinearSolverType* linearSolver = new g2o::LinearSolverCholmod<Block::PoseMatrixType>(); // 线性方程求解器
std::unique_ptr<Block::LinearSolverType> linearSolver ( new g2o::LinearSolverCholmod<Block::PoseMatrixType>());
//Block* solver_ptr = new Block ( linearSolver ); // 矩阵块求解器
std::unique_ptr<Block> solver_ptr ( new Block ( std::move(linearSolver))); // 矩阵块求解器
// 梯度下降方法，从GN, LM, DogLeg 中选
//g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( solver_ptr );
g2o::OptimizationAlgorithmLevenberg* solver = new g2o::OptimizationAlgorithmLevenberg ( std::move(solver_ptr));
5.编译问题
slambook-master/ch11/pose_graph_g2o_lie_algebra.cpp:77:67: error: no matching function for call to ‘Sophus::SO3<double>::SO3(const double&, const double&, const double&)’
77 | Sophus::SO3<double> ( update[3], update[4], update[5] ),
//TODO 有待解决
- slambook-master/ch12
1.编译问题
CMake Error at CMakeLists.txt:8 (find_package):
Could not find a configuration file for package "OpenCV" that is compatible
with requested version "3.1".
The following configuration files were considered but not accepted:
/usr/local/lib/cmake/opencv4/OpenCVConfig.cmake, version: 4.5.5
1.解决方案
将CMakeLists改为：
# opencv
#find_package( OpenCV 3.1 REQUIRED )
find_package( OpenCV REQUIRED )
2.编译问题
-- Build files have been written to: /mnt/d/slam14/slambook-master/ch12/build
make[2]: *** No rule to make target '/usr/local/lib/libDBoW3.a', needed by 'feature_training'. Stop.
make[2]: *** Waiting for unfinished jobs....
make[2]: *** No rule to make target '/usr/local/lib/libDBoW3.a', needed by 'loop_closure'. Stop.
make[2]: *** Waiting for unfinished jobs....
make[2]: *** No rule to make target '/usr/local/lib/libDBoW3.a', needed by 'gen_vocab'. Stop.
make[2]: *** Waiting for unfinished jobs....
[ 50%] Building CXX object CMakeFiles/gen_vocab.dir/gen_vocab_large.cpp.o
[ 50%] Building CXX object CMakeFiles/loop_closure.dir/loop_closure.cpp.o
[ 50%] Building CXX object CMakeFiles/feature_training.dir/feature_training.cpp.o
make[1]: *** [CMakeFiles/Makefile2:139: CMakeFiles/gen_vocab.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
make[1]: *** [CMakeFiles/Makefile2:87: CMakeFiles/feature_training.dir/all] Error 2
make[1]: *** [CMakeFiles/Makefile2:113: CMakeFiles/loop_closure.dir/all] Error 2
make: *** [Makefile:91: all] Error 2
2.解决方案
由于/usr/local/lib/中只有libDBoW3.so没有libDBoW3.a，将CMakeLists中libDBoW3.a改为libDBoW3.so即可，如下所示：
#set( DBoW3_LIBS "/usr/local/lib/libDBoW3.a" )
set( DBoW3_LIBS "/usr/local/lib/libDBoW3.so" )
#参考链接
高翔视觉slam十四讲第二版第十一章DBow3词库安装与出现的问题 - JavaShuo
- slambook-master/ch13
1.编译问题
[ 75%] Building CXX object CMakeFiles/pointcloud_mapping.dir/pointcloud_mapping.cpp.o
In file included from /usr/local/include/pcl-1.12/pcl/pcl_macros.h:74,
from /usr/local/include/pcl-1.12/pcl/impl/point_types.hpp:42,
from /usr/local/include/pcl-1.12/pcl/point_types.h:354,
from /mnt/d/slam14/slambook-master/ch13/dense_RGBD/pointcloud_mapping.cpp:9:
/usr/local/include/pcl-1.12/pcl/pcl_config.h:7:4: error: #error PCL requires C++14 or above
7 | # PCL requires C++14 or above
| ^~~~~
1.解决方案
在CMakeLists中添加
set( CMAKE_CXX_STANDARD 14)
2.编译问题
slambook-master/ch13/dense_RGBD/build# ./octomap_mapping
cannot find pose file
slambook-master/ch13/dense_RGBD/build# ./pointcloud_mapping
cannot find pose file
回退到源码目录，执行./build/pointcloud_mapping命令，如下：
slambook-master/ch13/dense_RGBD# ./build/pointcloud_mapping
Segmentation fault
//TODO有待解决，可能是由于UI界面显示问题导致的
3.编译问题
CMake Error at CMakeLists.txt:11 (find_package):
Could not find a configuration file for package "OpenCV" that is compatible
with requested version "3.1".
The following configuration files were considered but not accepted:
/usr/local/lib/cmake/opencv4/OpenCVConfig.cmake, version: 4.5.5
3.解决方案
将CMakeLists改为：
# opencv
#find_package( OpenCV 3.1 REQUIRED )
find_package( OpenCV REQUIRED )
4.编译问题
slambook-master/ch13/dense_monocular/dense_mapping.cpp:8:10: fatal error: sophus/se3.h: No such file or directory
8 | #include <sophus/se3.h>
4.解决方案
// for sophus
//#include <sophus/se3.h>
#include <sophus/se3.hpp>
5.编译问题
/mnt/d/slam14/slambook-master/ch13/dense_monocular/dense_mapping.cpp:402:34: error: ‘CV_GRAY2BGR’ was not declared in this scope
402 | cv::cvtColor( ref, ref_show, CV_GRAY2BGR );
5.解决方案
这是由于OpenCV版本更新导致的，用COLOR_GRAY2BGR替换掉CV_GRAY2BGR即可。
6.编译问题
slambook-master/ch13/dense_monocular/dense_mapping.cpp:51:15: error: type/value mismatch at argument 1 in template parameter list for ‘template<class _Tp, class _Alloc> class std::vector’
51 | vector<SE3>& poses
| ^
/mnt/d/slam14/slambook-master/ch13/dense_monocular/dense_mapping.cpp:51:15: note: expected a type, got ‘SE3’
/mnt/d/slam14/slambook-master/ch13/dense_monocular/dense_mapping.cpp:51:15: error: template argument 2 is invalid
/mnt/d/slam14/slambook-master/ch13/dense_monocular/dense_mapping.cpp:58:11: error: invalid use of template-name ‘Sophus::SE3’ without an argument list
58 | const SE3& T_C_R,
6.解决方案
//TODO
将SE3修改为SE3d,例如：
//using Sophus::SE3;
using Sophus::SE3d;
7.编译问题
/mnt/d/slam14/slambook-master/ch13/dense_monocular/dense_mapping.cpp:351:25: error: ‘using SE3d = class Sophus::SE3<double>’ {aka ‘class Sophus::SE3<double>’} has no member named ‘rotation_matrix’; did you mean ‘rotationMatrix’?
351 | Vector3d f2 = T_R_C.rotation_matrix() * f_curr;
| ^~~~~~~~~~~~~~~
| rotationMatrix
7.解决方案
将‘rotation_matrix’修改为‘rotationMatrix’
8.编译问题
[100%] Linking CXX executable dense_mapping
/usr/bin/ld: CMakeFiles/dense_mapping.dir/dense_mapping.cpp.o: in function `Sophus::SO3Base<Sophus::SO3<double, 0> >::normalize() [clone .part.0]':
dense_mapping.cpp:(.text+0x171): undefined reference to `fmt::v8::vprint(fmt::v8::basic_string_view<char>, fmt::v8::basic_format_args<fmt::v8::basic_format_context<fmt::v8::appender, char> >)'
/usr/bin/ld: CMakeFiles/dense_mapping.dir/dense_mapping.cpp.o: in function `updateDepthFilter(Eigen::Matrix<double, 2, 1, 0, 2, 1> const&, Eigen::Matrix<double, 2, 1, 0, 2, 1> const&, Sophus::SE3<double, 0> const&, cv::Mat&, cv::Mat&)':
8.解决方案
在CMakeLists中添加fmt库链接，如下所示：
#target_link_libraries( dense_mapping ${THIRD_PARTY_LIBS})
target_link_libraries( dense_mapping ${THIRD_PARTY_LIBS} fmt)
9.编译问题
slambook-master/ch13/dense_RGBD# pcl_viewer map.pcd
2022-02-09 00:18:10.525 ( 0.158s) [ 7B4A0800] vtkContextDevice2D.cxx:32 WARN| Error: no override found for 'vtkContextDevice2D'.
The viewer window provides interactive commands; for help, press 'h' or 'H' from within the window.
> Loading map.pcd 2022-02-09 00:18:10.964 ( 0.597s) [ 7B4A0800]vtkXOpenGLRenderWindow.:1212 ERR| vtkXOpenGLRenderWindow (0x5606ef7f5b80): bad X server connection. DISPLAY=
Aborted
9.解决方案
//TODO有待解决掉其他问题后一并解决
应该是由于wsl没有安装UI界面，导致编译的程序调用界面显示时不能显示所致。
其他说明
slambook-master/ch13/dense_monocular编译完后，需要下载remode_test_data.zip并解压到源码目录（下载链接http://rpg.ifi.uzh.ch/datasets/remode_test_data.zip）。然后进入slambook-master/ch13/dense_monocular/build目录，执行./dense_mapping ../test_data即可运行。
扫一扫
1011
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
