Open eleanorchienqr opened 2 years ago
Thank you for your comment. I agree with you, according to my evaluation results orb_slam3 is more accurate regarding V1_02. Your question can be answered by reading my paper which is currently under review and is not yet publicly available. The extra use of line features is not inextricably connected to higher precision. However, orb_line_slam dominates orb_slam3 on EuRoC dataset in total.
I test the algorithm on my dataset, and it also verifies that the rmse result of ORB-LINE-SLAM is higher than ORB-SLAM3.
Where is your paper published
Analysing evaluation on dataset V1_02 of EuRoC, the rmse result of ORB-LINE-SLAM is higher than ORB-SLAM3. But ORB-LINE-SLAM uses extra line features to track and estimate poses, which means higher precision and lower rmse. I'm wondering why the higher rmse occurs? Thank you!