g2o指针报错

g2o指针报错

1.问题描述一

error: no matching function for call to ‘g2o::BlockSolver<g2o::BlockSolverTraits<-1, -1> >::BlockSolver(SlamLinearSolver*&)

并且提示

std::unique_ptr

指针原因，可能是重复引用
参考链接
https://blog.csdn.net/robinhjwy/article/details/78084210.
https://blog.csdn.net/lun55423/article/details/109036995.
https://www.cnblogs.com/xueyuanaichiyu/p/7921382.html.
https://www.cnblogs.com/serser/p/7815588.html.

进行代码改动：
初始代码：

G2oSolver::G2oSolver()
{
  // 第1步：创建一个线性求解器LinearSolver
  SlamLinearSolver *linearSolver = new SlamLinearSolver();
  linearSolver->setBlockOrdering(false);
  // 第2步：创建BlockSolver。并用上面定义的线性求解器初始化
  SlamBlockSolver *blockSolver = new SlamBlockSolver(linearSolver);
  // 第3步：创建总求解器solver。并从GN, LM, DogLeg 中选一个，再用上述块求解器BlockSolver初始化
  g2o::OptimizationAlgorithmLevenberg *solver = new g2o::OptimizationAlgorithmLevenberg(blockSolver);
  // 第4步：创建稀疏优化器（SparseOptimizer）
  mOptimizer.setAlgorithm(solver);
}

改动后代码：

G2oSolver::G2oSolver()
{
  // 第1步：创建一个线性求解器LinearSolver
  SlamLinearSolver *linearSolver = new SlamLinearSolver();
  linearSolver->setBlockOrdering(false);
  // 第2步：创建BlockSolver。并用上面定义的线性求解器初始化
  SlamBlockSolver *blockSolver=new SlamBlockSolver(std::unique_ptr<SlamLinearSolver>(linearSolver));
  // 第3步：创建总求解器solver。并从GN, LM, DogLeg 中选一个，再用上述块求解器BlockSolver初始化
  g2o::OptimizationAlgorithmLevenberg *solver = new g2o::OptimizationAlgorithmLevenberg(std::unique_ptr<SlamBlockSolver>(blockSolver));
  // 第4步：创建稀疏优化器（SparseOptimizer）
  mOptimizer.setAlgorithm(solver);
}

2.问题描述二

fatal error: cs.h: No such file or directory

问题出现在
/usr/local/include/g2o/solvers/csparse/csparse_extension.h的

#include <cs.h>

安装
CSparse

sudo apt-get install libsuitesparse-dev

找到头文件

/usr/include/suitesparse/cs.h

在CMakeLists.txt中包含头文件

include_directories("/usr/include/suitesparse") 
target_link_libraries(${PROJECT_NAME}_karto_slam_node
  ${catkin_LIBRARIES}
  ${Boost_LIBRARIES}
  sba
  ${CERES_LIBRARIES}
  gtsam

  ${G2O_LIBRARIES}
  ${SUITESPARSE_LIBRARIES}
)

参考链接：
https://blog.csdn.net/weixin_43011182/article/details/118878559.
https://blog.csdn.net/weixin_44436677/article/details/106095485.