Feature-based, Direct, and Deep Learning Methods of Visual Odometry

Поделиться
HTML-код
  • Опубликовано: 28 дек 2024

Комментарии • 9

  • @slam_slam_
    @slam_slam_ 4 года назад +4

    Great talk!
    At @34:52, I think the 5-point algorithm is actually not proposed recently (i.e. in 2014) - in fact it's proposed quite a long time ago in 2004!

    • @yafeihu5098
      @yafeihu5098 2 года назад +2

      Right, the paper I mentioned in the slides is the improved version. Thank you for mentioning that.

  • @LuigiFreda
    @LuigiFreda 3 года назад +5

    Great talk! Thanks for suggesting pyslam :-)

    • @yafeihu5098
      @yafeihu5098 2 года назад

      Thank you for releasing such a good project!

  • @tyoung96
    @tyoung96 2 года назад

    Thanks for great talk.
    I have question about tutorial code.
    I want to compare the TUM dataset ground trajectory and the estimated trajectory in 'vo.py' and visualize it, but how do I know the scale value?
    It doesn't look like similar when I test it.

  • @karanbirchahal3268
    @karanbirchahal3268 2 года назад

    What a banger

  • @rabomeister
    @rabomeister 2 года назад +2

    Next time consider removing the Zoom pics

  • @고금명-m8h
    @고금명-m8h Год назад

    学长你好,我是研一新生,看了学长的视频,还是感觉十分的晦涩,是不是因为我的数学底子太差了?orz