Ideally you should get the original orthophoto used to create the textures of the mesh. I can't think of an FME workflow that would let you do this. But it's an interesting problem for sure. I shall follow this thread with interest. 😏
Ideally you should get the original orthophoto used to create the textures of the mesh. I can't think of an FME workflow that would let you do this. But it's an interesting problem for sure. I shall follow this thread with interest. 😏
Getting the original orthophoto would be ideal, but unfortunately not possible.
I tried converting the mesh to a pointcloud and then using the imageRasterizer and while that mostly worked, there were a few surprises, particularly in the tree at the lower right.
I might need to try and filter the pointcloud to only the highest z or something.
Getting the original orthophoto would be ideal, but unfortunately not possible.
I tried converting the mesh to a pointcloud and then using the imageRasterizer and while that mostly worked, there were a few surprises, particularly in the tree at the lower right.
I might need to try and filter the pointcloud to only the highest z or something.
Ohhh of course, what an excellent idea. Yeah very strange in the corner there, strange that it's very much in a grid form. I've has some issues when converting point clouds to raster, I've found that snapping the points to a grid and then getting the max z for each new grid point works well - there were some tricks in a couple of point cloud articles that helped with this. The data I worked with though was not RGB but height values so didn't matter to much about how it looked.
I had the same idea and the same problems.
I solved this problem by using the PointCloudSorter and sorting the Z component in ascending order. That resolved the problem in my case.