Hi there,
I am trying to produce high quality ortho photos of a building facade, is it not possible to achieve this with capturing reality software? The ones I’ve created so far are very pixelated.
Thanks
Warren
Hi Warren,
The Orthomosaic from RealityCapture differs from other photogrammetry processing software in that RC’s Ortho is a top down view of the textured 3D model (or horizontal in your case as it’s a building facade).
Other software packages will usually create a mesh (triangulated point cloud in the form of a DTM) on which they project the original images.
Therefor if you want to create a high resolution Ortho with RC you will have to create a textured 3D model with the highest texture resolution. Then you set your ortho bounding box accordingly and generate an Ortho with the same resolution as your textured model.
Thanks Jasper,
That’s exactly what I’ve been doing, the tiff file generated is huge 3.20gb but the quality isn’t that great, I can’t view the textured model as it has too many triangles.
I went through an unwrap procedure to up the quality of the mesh but it’s still saying 11% texture quality.
Seem to be going round in circles.
Hi Warren,
Could you share a small screenshot of your unwrap settings?
What you want to do is set the unwrap style to Fixed Texel Size and then set this fixed value to “Optimal Texel Size”
This optimal texel size value refers to the average pixel resolution across all photos. This means if all your captured images were taken at roughly the same distance to your building facade then this value will be true for the average value. This might be different if you have a lot of variation in your images in terms of distance to the object and thus their pixel resolution.
Any good?
Hi Warren,
You’ll have to click on the small + sign near the bottom of your screenshot where it says “+Unwrap Parameters”
There you can set the style to fixed texel size and put it on optimal.
Once the unwrap is finished, press on “Texture” again to apply the new textures (might take some time)
Then go to the Ortho Projection tool in the Reconstruction tab and place your ortho bounding box accordingly.
In the “Ortho Pixel Size” value box, fill in the exact same value that’s filled in at the Optimal Texel Size in the unwrap tool.
Then press Render. You should end up with a high resolution rendered image that has the same resolution as your textured model.
Thanks, giving it a go now, I’ll keep you posted!
By the way does RC have it’s own viewer? I’ve been trialling context capture and it has it’s own viewer which seems to only load the triangles it needs, super fast and very high resolution.
Okay, thanks for your helps that’s much better.