Texture quality and optimal settings

Hi I’m a fairly new user to RC and its been really awesome for the things it can do. I’m mainly doing Medium scale architecture and site photogrammetry through drone imagery combined with laser scans and SLR imagery. 

Im having a recurring issue with texture quality however. Textures would appear blurry or not as detailed as expected given the clarity and amount of images taken. 

This is an example of a site captured with 1004 images from a Mavic Pro 2, 130 LSP and 60 slr images. unwrapped with fixed textel size at 8k gives utilization of 97% and 108% texture quality, does this mean its already at its maximum texture quality?

 

These are rendered from RC and shows quite blurry for some areas which makes me question if it is an alignment issue but the general mesh looks aligned especially as lsp and control points were used. 

I also changed the weighting of the images so those closest to the ground was weighted at 1.0 while higher/ overall imagery was weighted at 0.5. 

Some of the Imagery: 

Just wondering if i am doing incorrect steps in alignment or texturing or if this is the expected quality and my expectations are too high. 

Also, welcome any tips on downstream post-process programs to optimize the mesh or textures for game engine type visualisation. 

Appreciate any help on this :) 

 

 

Hi Isaac,

            Laser scans are great for alignment and meshing, but you will probably get better results if you disable them for texturing. Some of the images suggest you might have some misalignments, but it’s better to look at the point cloud to look for “ghosting”, for instance two almost parallel lines of tie points along the ridge of that roof. See if you can add some control points in these areas, if they are not aligned. Also 60 seems like a very low amount of DSLR images. You might find this webinar useful RealityCapture Webinar: Combining Z+F laser scan with images inside of RealityCapture or this more in depth one RealityCapture Free Webinar: Advanced workflow for a combination of images and laser scans