4 ms·

This is true because they use cameras and camera lack depth information despite it being used with “state of the art AI”. This would not be an issue if they hav

This is true because they use cameras and camera lack depth information despite it being used with “state of the art AI”. This would not be an issue if they have any real depth based measurement system like a LiDAR.

jeffwass 4y ago
How hard is it to use multiple cameras to gauge depth perception, as per human binocular vision? (Serious question, I know nil about machine vision)
- refulgentis 4y ago
  It's well understood to be impossible / not there yet, the argument then shifts to "but ML is awesome so it could happen?" to which the correct response is "anything is possible but we're nowhere close"
- Gaessaki 4y ago
  In some ways, it can be better than lidar. The colour signal can be perfectly matched with the depth signal to provide further context. This is particularly useful for tracking moving objects. Vision-based systems also have a far greater resolution, especially when additional cameras are added. But having both would be best. The drawbacks of camera are additional compute being necessary for stitching and the risk of occlusions occurring due to something like adverse weather conditions. It also only offers a calculated estimation rather than a true distance measurement.
- syntaxing 4y ago
  It’s extremely difficult, the biggest issue is that any vision based depth system is so damn noisy.
- hnburnsy 4y ago
  Do Teslas have 2 main forward cameras? I know that Subarus do as you can see them https://www.thetruthaboutcars.com/2012/03/subaru-uses-stereo-3d-tech-in-new-eyesight-adas/ https://www.thetruthaboutcars.com/2012/03/subaru-uses-stereo...
- tuatoru 4y ago
  Related, why is rangefinding like on rangefinder cameras not usable? 1. videos: https://www.lensangle.com/how-do-rangefinder-cameras-work/ https://www.lensangle.com/how-do-rangefinder-cameras-work/ 2. a photo or two: https://www.howtogeek.com/365801/what-is-a-rangefinder-camera/ https://www.howtogeek.com/365801/what-is-a-rangefinder-camer...

jeffwass 4y ago

How hard is it to use multiple cameras to gauge depth perception, as per human binocular vision? (Serious question, I know nil about machine vision)

refulgentis 4y ago
It's well understood to be impossible / not there yet, the argument then shifts to "but ML is awesome so it could happen?" to which the correct response is "anything is possible but we're nowhere close"
Gaessaki 4y ago
In some ways, it can be better than lidar. The colour signal can be perfectly matched with the depth signal to provide further context. This is particularly useful for tracking moving objects. Vision-based systems also have a far greater resolution, especially when additional cameras are added. But having both would be best. The drawbacks of camera are additional compute being necessary for stitching and the risk of occlusions occurring due to something like adverse weather conditions. It also only offers a calculated estimation rather than a true distance measurement.
syntaxing 4y ago
It’s extremely difficult, the biggest issue is that any vision based depth system is so damn noisy.
hnburnsy 4y ago
Do Teslas have 2 main forward cameras? I know that Subarus do as you can see them https://www.thetruthaboutcars.com/2012/03/subaru-uses-stereo-3d-tech-in-new-eyesight-adas/ https://www.thetruthaboutcars.com/2012/03/subaru-uses-stereo...
tuatoru 4y ago
Related, why is rangefinding like on rangefinder cameras not usable? 1. videos: https://www.lensangle.com/how-do-rangefinder-cameras-work/ https://www.lensangle.com/how-do-rangefinder-cameras-work/ 2. a photo or two: https://www.howtogeek.com/365801/what-is-a-rangefinder-camera/ https://www.howtogeek.com/365801/what-is-a-rangefinder-camer...