Hi All
New to the forum here, and relatively new to Photoscan. I ran a test using a GoPro on a painters pole to see how well I could get it to work on a couple of large statues either side of a doorway. After earlier trials I decided on just using the normal FOV on the GoPro rather than the wide, but I still got a lot of noise around the edges. After looking around at other photogrammetry projects I decided to do a test on a simpler subject (brick wall with bluestone cobblestones in a laneway to try and see if Photoscan could calculate a reasonable set of distortion parameters. Whilst not a precise mehodology, you can at least assess the quality of the result by the flatness of the wall and the amount of noise away from the surface of the wall and ground.
- 3 rows of photos so that the edges of one row overlapped the centre of the next.
- Align images in Photoscan and look at the calculated camera calibration.
- Repeat 3 time with camera roll angles of -90, 0 and 90°
The resulting point cloud looked pretty good compared to my previous tests so I saved the lens calibration. I then loaded the images of the statues and re-ran the workflow with the new lens calibration. The resulting model was much better although noise is still a problem.
Not sure that I'd use the GoPro for aerial work though. you'd have to fly pretty low to get decent resolution, but that depends on the application I guess.