Lack of Extrinsic Parameter from Mocap to Lidar

[sjtu-visys]

Lack of Extrinsic Parameter from Mocap to Lidar

[yuhui-xie]

As mentioned in the paper, the ground truth trajectories of Room and Roomdark sequences are obtained by Mocap. In fact, Mocap paremeters are not necessary, because when you evaluate your SLAM result, the "-a" command of EVO will align the GT with your trajectories. We do not have to specially calibrate the extrinsics between LiDAR and Mocap.

It means that actually the EVO will estimate the extrinic in the alignment process? Like hand-eyes calibration between two trajectory?

[sjtuyinjie]

As mentioned in the paper, the ground truth trajectories of Room and Roomdark sequences are obtained by Mocap. In fact, Mocap paremeters are not necessary, because when you evaluate your SLAM result, the "-a" command of EVO will align the GT with your trajectories. We do not have to specially calibrate the extrinsics between LiDAR and Mocap.

[oym1994]

Got it! Actually we want to build the map with Room datasets with ground-truth poses and locate with the Roomdark datasets. In this case, the absolute ground-truth pose is necessary. But if the extrinsic parameter is not available, just ignore it. Anywaym thank you for your kind response!

[sjtuyinjie]

OK. If any other questions, feel free to propose issues.