I’ve been doing some tests with reality-scan because of its good reputation. I used Map Wizard to create an orthophoto and a model, however, the results are so bad that I suspect something strange has happened.
The orthophoto has many artifacts, jagged edges, duplicate areas, etc.
The model, on the other hand, has a very poor texture, and although I’ve rebuilt the texture several times, I still haven’t gotten anything decent.
It really depends on the used settings. Such orthos were created using mosaicing. Such orthos can be modified.
Also, the view in RealityScan 3D view is reduced. To see the proper resolution you need to provide the rendering.
What do you mean by the orthophoto being modifiable? Are these results normal? I used Map Wizard to generate it; would you like me to provide the configuration parameters?
Regarding the model, how do I provide the rendering? Do you mean exporting it and viewing it in another program?
There is an option how to modify the created ortho. It is called enhance mosaic: https://www.youtube.com/watch?v=JxNwvpNsJUI. There is also a tutorial about Image mosaicing in general: https://www.youtube.com/watch?v=fMWuHA7gh-Q.
For such projects the usage of Map Wizard can be tricky, as you don’t have such control over the process (you only inject the settings and wait for the result). For most cases it would be better to provide the project step by step to see how the application actually works. Then you can use the working settings in the wizard.
But is this normal? The images I uploaded are just a few examples, but the artifacts I’ve observed are widespread throughout the entire production. I’ve watched the videos, but manually editing these errors would be impractical since they aren’t isolated artifacts; they appear everywhere. What recommended settings should I use? What could be happening?
Regarding the model, is the texture issue normal? It looks like it’s been rebuilt at a very low resolution. What could have happened?
For ortho mosaicing it can happen. Mostly for building. It depends which images are selected for such part of the building. Therefore it needs to be polished manually, as the selection is automated. Also, the tool is still in beta, so it would be better to use ordinary ortho generator.
About the model, it really depends on the used settings. This looks more like as just coloring was used or some simplification without reprojection. Hard to say just from the images. Also, if that is a very big textured model, there are some restrictions to show the whole texture resolution in the 3D view. To see the proper textures, you need to render the view.
