When evaluating an algorithm in the VIO SLAM system, the evaluation is performed using benchmark data. For most evaluation results, translation-error and rotation-error are calculated and returned. Which of the two results is more important? I personally think translation-error is more important, but there is no obvious reason. So I ask a question. I would be grateful if you could give me any advice.
1.4m articles
1.4m replys
5 comments
56.9k users