The original point is diffrent two much between gt and algo when use evo_** gt_file algo_out_flie -p -a

[jiajiaandWu]

(pythonvo) PS D:\YOLOPoint-VO\dataprocess> evo_traj tum --ref E:\results_yoloVSmomo\groundtruthTUM_100hz.txt E:\results_yoloVSmomo\vins_result_loop.txt --plot_mode=xy -a -p -v

Loaded 1066 stamps and poses from: E:\results_yoloVSmomo\vins_result_loop.txt
Loaded 29491 stamps and poses from: E:\results_yoloVSmomo\groundtruthTUM_100hz.txt

Found 1066 of max. 1066 possible matching timestamps between...
reference
and: E:\results_yoloVSmomo\vins_result_loop.txt
..with max. time diff.: 0.01 (s) and time offset: 0.0 (s).

Aligning E:\results_yoloVSmomo\vins_result_loop.txt to reference.
Aligning using Umeyama's method...
Rotation of alignment:
[[ 0.87119358 0.490776 0.012675 ]
[-0.49093102 0.87104098 0.01656365]
[-0.0029114 -0.02065269 0.99978247]]
Translation of alignment:
[-57.63636653 18.41824369 -10.495273 ]
Scale correction: 1.0

name: vins_result_loop
infos:
duration (s) 294.5486240386963
nr. of poses 1066
path length (m) 1632.3046841376645
pos_end (m) [-175.805577 270.4669114 -4.21548744]
pos_start (m) [-56.44302709 16.53013172 -10.41201644]
t_end (s) 1605247735.4236526
t_start (s) 1605247440.8750286

name: groundtruthTUM_100hz
infos:
duration (s) 294.90000009536743
nr. of poses 29491
path length (m) 1507.02978251964
pos_end (m) [-156.21134 250.127887 -7.8 ]
pos_start (m) [ 2.068028 -2.597068 -0.3 ]
t_end (s) 1605247735.7
t_start (s) 1605247440.8

[MichaelGrupp]

Your trajectories don't match well. Without any context one can only guess what you want to do.

Please provide a proper description if you have a specific question about evo.