Ceres Solver是一个开源的C++库，用于建模和解决大型、复杂的优化问题。它是一个成熟、功能丰富且高性能的库，自2010年以来一直在Google生产中使用。最新发布版本为2.1.0,license为BSD,它支持在Windows、Linux、Mac、Android、iOS上编译，源码地址：https://ceres-solver.googlesource.com/ceres-solver
Ceres Solver可用于解决两类问题：
(1).具有边界约束的非线性最小二乘(Non-linear Least Squares)问题;
(2).一般无约束优化问题。
Ceres Solver依赖项：
(1).必须的：CMake、Eigen、glog、gflags；
(2).可选的：SuiteSparse、BLAS and LAPACK、CUDA。
Ceres Solver在Ubuntu 20.04上编译：eigen, glog, gflags, ceres-solver都在/home/spring/Soft目录下
(1).clone eigen源码：这里使用3.3.7版本，依次执行如下命令：
git clone https://gitlab.com/libeigen/eigen.git
cd eigen
git checkout 3.3.7编译eigen，在eigen目录下执行build_eigen.sh，其内容如下：
#! /bin/bash
if [[ ! -d "build" ]]; then
mkdir build
cd build
else
cd build
fi
cmake \
-DCMAKE_CXX_FLAGS=-fPIC \
-DCMAKE_C_FLAGS=-fPIC \
-DCMAKE_BUILD_TYPE=Release \
-DCMAKE_INSTALL_PREFIX=../install \
..
make -j2
make install(2).clone glog源码：这里使用v0.4.0版本，依次执行如下命令：
git clone https://github.com/google/glog.git
cd glog
git checkout v0.4.0编译glog，在glog目录下直接执行build_glog.sh，其内容与build_eigen.sh完全一致。
(3).clone gflags源码：这里使用v2.2.2版本，依次执行如下命令:
git clone https://github.com/gflags/gflags.git
cd gflags
git checkout v2.2.2编译gflags,在gflags目录执行build_gflags.sh，其内容与build_eigen.sh完全一致。
(4).clone Ceres Solver源码：这里使用1.14.0版本，依次执行如下命令:
git clone https://ceres-solver.googlesource.com/ceres-solver
cd ceres-solver
git checkout 1.14.0编译ceres-solver,在ceres-solver目录执行build_ceres-solver.sh，其内容如下：
#! /bin/bash
if [[ ! -d "build" ]]; then
mkdir build
cd build
else
cd build
fi
path=/home/spring/Soft
cmake \
-DCMAKE_CXX_FLAGS=-fPIC \
-DCMAKE_C_FLAGS=-fPIC \
-DCMAKE_BUILD_TYPE=Release \
-DCMAKE_INSTALL_PREFIX=../install \
-DEigen3_DIR=${path}/eigen/install/share/eigen3/cmake \
-Dglog_DIR=${path}/glog/install/lib/cmake/glog \
-Dgflags_DIR=${path}/gflags/install/lib/cmake/gflags \
..
make -j2
make installCeres Solver调用：在/home/spring/Soft目录下新建test目录,此目录内文件内容依次如下
(1).main.cpp:
#include <iostream>
#include <glog/logging.h>
#include <ceres/ceres.h>
using namespace ceres;
struct CostFunctor {
template <typename T>
bool operator()(const T* const x, T* residual) const {
residual[0] = 10.0 - x[0];
return true;
}
};
int main(int argc, char** argv)
{
// reference: http://ceres-solver.org/nnls_tutorial.html
google::InitGoogleLogging(argv[0]);
// The variable to solve for with its initial value.
double initial_x = 5.0;
double x = initial_x;
// Build the problem.
Problem problem;
// Set up the only cost function (also known as residual). This uses
// auto-differentiation to obtain the derivative (jacobian).
CostFunction* cost_function = new AutoDiffCostFunction<CostFunctor, 1, 1>(new CostFunctor);
problem.AddResidualBlock(cost_function, nullptr, &x);
// Run the solver!
Solver::Options options;
options.linear_solver_type = ceres::DENSE_QR;
options.minimizer_progress_to_stdout = true;
Solver::Summary summary;
Solve(options, &problem, &summary);
std::cout << summary.BriefReport() << "\n";
std::cout << "x : " << initial_x << " -> " << x << "\n";
return 0;
}(2).build.sh:
#! /bin/bash
if [[ ! -d "build" ]]; then
mkdir build
cd build
else
cd build
fi
path=/home/spring/Soft
cp ${path}/glog/install/lib/libglog.a .
cp ${path}/gflags/install/lib/libgflags.a .
cp ${path}/ceres-solver/install/lib/libceres.a .
cmake ..
make(3).CMakeLists.txt:
cmake_minimum_required(VERSION 3.15)
project(test_ceres-solver)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fopenmp")
include_directories(
${CMAKE_CURRENT_SOURCE_DIR}/../eigen/install/include/eigen3
${CMAKE_CURRENT_SOURCE_DIR}/../glog/install/include
${CMAKE_CURRENT_SOURCE_DIR}/../gflags/install/include
${CMAKE_CURRENT_SOURCE_DIR}/../ceres-solver/install/include
)
add_executable(${CMAKE_PROJECT_NAME} ${CMAKE_CURRENT_SOURCE_DIR}/main.cpp)
target_link_directories(${CMAKE_PROJECT_NAME} PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/build)
target_link_libraries(${CMAKE_PROJECT_NAME} ceres glog gflags pthread)执行：
首先执行：./build.sh
然后执行：./build/test_ceres-solver ,结果如下图所示：
