Wind Turbine 3D model

I’m using images from the frames of a video of a Wind Turbine Generator captured through drone.
The frames doesn’t get stitched properly nor the point cloud is good. I also checked with removal of background and splitting the datasets for different parts. That seemed to work.

  1. How can I process the frames of the video.
  2. How can I merge the point cloud obtained in a single projects from different tasks.
  3. What processing options I’ve to select in order to get good results in this case.
  1. What resolution and framerate is the video? How often are you sampling from the video?
    1A) Are you getting a lot of sky? If so, you should mask it
  2. Likely will need to use something like CloudCompare to merge different point clouds
  3. Really hard to say without knowing the data and what you think fitness for purpose is
1 Like

Frame Rate - 29.97 frames/second
Resolution - 2688*1512

  1. I tried with different samplings. 1/10,1/50,1/100 and No sampling as well.
  2. A proper 3D model of a wind turbine from drone captured data will be considered as good fit.kang225

4MP is a bit iffy. I’ve had trouble with mobile/small sensors below 8MP. You may need to pre-process those frames.

  1. Sampling depends upon how fast you’re moving and if you can keep good overlap at your sample rate. If you can spare the RAM and processing time, I guess no sampling is pretty safe.
  2. You may need to mask the sky at the least.
1 Like