数据集对应的位姿格式

1.kitti

More infos: The KITTI Vision Benchmark Suite

 2.euroc

 3.tum

More infos: Computer Vision Group - File Formats

3.1输出tum格式的相关代码

// 添加定义

std::string RESULT_PATH;
int init_flag=true;
float tum[8];
double first_time;

// 相关输出代码
if (init_flag==true)	
        {
            first_time = laserOdometry->header.stamp.toSec();
            init_flag=false;
        }

        tum[0] = laserOdometry->header.stamp.toSec() - first_time;
        tum[1] = t_w_curr.x();
        tum[2] = t_w_curr.y();
        tum[3] = t_w_curr.z();
        tum[4] = q_w_curr.x();
        tum[5] = q_w_curr.y();
        tum[6] = q_w_curr.z();
        tum[7] = q_w_curr.w();


        std::ofstream foutC(RESULT_PATH, std::ios::app);

        foutC.setf(std::ios::scientific, std::ios::floatfield);
        foutC.precision(18);
        for (int j = 0; j < 8; ++j)
                {
                if(j==7)
                {
                    foutC <<tum[j]<< std::endl ;	
                }
                else
                {
                    foutC <<tum[j]<< " " ;
                }              
                }
        foutC.close();
}

 4.格式转换

5.kitti转tum方法：https://github.com/MichaelGrupp/evo/blob/master/contrib/kitti_poses_and_timestamps_to_trajectory.py

python kitti_poses_and_timestamps_to_trajectory.py ./05/05.txt ./05/times.txt ./05/tum_05_gt.txt

注意：

kitti_poses_and_timestamps_to_trajectory.py中import的部分包来自于evo程序中，所以必须将该py文件放置在evo文件夹中。

来源：Formats · MichaelGrupp/evo Wiki · GitHub