1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

LIDAR for the masses

Discussion in 'Racer' started by radome, Nov 5, 2012.

  1. radome


  2. Mr Whippy

    Mr Whippy

    Hehe that is pretty interesting.

    I guess what really matters now is a scanning head systems where the scanner can accurately determine orientation of the head, log it and log the range values.

    Then you can start to build up a cloud of points.

    In practice there isn't much to a laser scanner. Usually what they do is spin the head about the pitch axis at a known rate, and then rotate the head about the vertical axis at a known rate.
    I guess as long as those rotational devices are consistent in operation and/or there is error correction designed in somehow, then rebuilding your captured ranges into a cloud is just some maths.

    Problem is having a laser bright enough for outdoors that a conventional camera will see. I think this is where pro scanners have a specific detector so that they can 'see' the relatively low energy spot in bright normal light conditions. That way you avoid blinding people around the scanner :D

    Also some will scan for colour too, so take a nice panoramic image for reference (usually apply the 'seen' colour to the point cloud points)

    I believe that commercial scanners also push blacks further away due to the wavelength changing, so perhaps they look at the wavelength of reflected light for determining ranging too?!

    This method of measuring the offset in theory works better as I can't see much wrong with it. Obviously eventually the resolution of the camera will be a source for your error as the spot moves within a pixels size on the sensor...
    However you could easily cover that issue by utilising a variable field of view lens, so you could scan with a wide lens, dump data beyond 20m say, then scan again with a zoom lens, and dump data within 20m.


    All very interesting stuff :D