Aerial Triangulation help

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • Frank Mullins
    3Dflower
    • Apr 2018
    • 5

    Aerial Triangulation help

    Hi All,

    I have problem. Could I ask what settings I should use to make a model from aerial photos ?

    These aerial photos are professional photos standard for making dems or othophotos, taken from the plane by pro camera.

    I know what is focal length for them, but here in Zephyr it should be adjusted in pixels, not in mm. What should be settings for k1, k2, k3, p1, p2 ? I attach an info file I got with images.

    I should add it is high mountain area. When I use standard (default) Zephyr settings, all the valley bottom areas are done OK, but all the summits are blank (no correlation at all for these areas). As if the areas, which are in big distance from camera are done OK, but these, which are closer to camera are not done at all.

    Help.
    Attached Files
  • Andrea Alessi
    3Dflow Staff
    • Oct 2013
    • 1335

    #2
    Hello Frank,

    focal lenght is just one of the intrisics parameter of a camera. Zephyr is completetly autocalibrated but you can make a calibration dataset if you wish. In most cases, you shouldn't have to do that unless you have a very high distortion and your dataset is suffering from drifting.

    From what you say, i think your issue is not related to calibration. If the summits seems "sliced" it simply may be a bounding box that is too small. Can you maybe post a screenshot of your issue?

    Comment

    • Frank Mullins
      3Dflower
      • Apr 2018
      • 5

      #3
      Hello,

      It is not an issue of too small bounding box.

      The summits are just correlating badly. I attach two screenshots: first with standard options with no Control Points; and the second, where I put 3 CPs. Control Points helped, but the quality of summits is still poor.

      Comment

      • Andrea Alessi
        3Dflow Staff
        • Oct 2013
        • 1335

        #4
        Hi Frank,

        any chance you could share the dataset with us? This way i can give you better advice on how to proceed to obtain better results.

        Comment

        Working...