Hi all,
I have a large number of images taken from a airplane with a GoPro Hero 7 fixed at the wing. I followed the ortho mapping workflow process for drone images with ArcGIS pro 2.4 and the result are very disappointed. Roads are very wavy and errors are huge, specially with the z coordinate (more than 20 m).
I know factors affecting the quality are lens' distortion and inaccuracy of the center of the image because the position of the plane. To see how product will look like and to learn the workflow I selected 11 images that covers rural and urban areas which, apparently, are not very leaned. I checked differences between the center of the image and the GPS location and the error is about 20-30 m. The images overlaps around 80% or more. In addition, I used 170 CGPs with coordinates obtained from Esri Imagery.
The images were taken with a GoPro Hero 7 camera in wide mode but either the fish-eye distortion is important. I've seen there is no parameters for this camera so I tried to input parameters manually, but I weren't able to find the correction parameters for this model. Then I used as camera model in Ortho Mapping the default for GoPro Hero 4 in the test I carried out. I know the quality of the input data is not the best desirable, but the distortion of the mosaic is really bad.
I've tried an alternative. First geo-referencing the images automatically and clipped them to have the same cols and rows and cell size. Then I create a mosaic dataset manually and I used it to make the workspace. In camera settings I set focal distance and pixel size so they match with the cell size of my georeferenced images. The mosaic looks better although errors are still high. One problem with the alternative approach is that the fails to generate the adjustment report.
Does anyone know the correct parameters for GoPro Hero 7 or has a different way to improve the quality of the ortho-mosaic?
Thank you,
Alfonso
