NODAR for Vehicle Vision - A Software Based Non-AI Approach

แชร์
ฝัง
  • เผยแพร่เมื่อ 21 ธ.ค. 2024

ความคิดเห็น •

  • @dragonlee2024
    @dragonlee2024 ปีที่แล้ว

    I have learned a lot from this podcast. Thank you!

  • @_Tidal_
    @_Tidal_ ปีที่แล้ว

    The core idea, that a wide baseline system is necessary as it allows for longer range depth estimates, seems a bit particular.
    Humans have very small baselines, about 5cm to 8cm. This is actually even smaller than those found on the already deployed stereo-based systems of Subarus/Mercedes. Nevertheless, as humans, we seem to be able to manage just fine - be it truckers or car drivers. So why push this particular angle? Depth perception may be important, but how important is it really? People who play racing car games/simulators on 2D TVs & monitors seem to be able to do so just fine without any need for depth. They have even demonstrated that those who played F1 racing simulators at the top-level were able to do relatively well when placed into actual F1 cars on real tracks despite never having previously driven the real thing.