export to meshlab - what is the best way to get a jpg of the model?

[Ethork]

I'm exporting the textured model to meshlab in order to get a jpg.  I'm exporting a obj file and inporting it into meshlab.
(Note that when you export the .obj file, you can't have any spaces in the file name.)
The results are better - higher resolution  - than exporting a pdf from agisoft, then converting to jpg in photoshop.
But the .obj file, when opened in meshlab, is lower resolution than the original agisoft model - it looks smooshy compared to the the crisp delineation of the agisoft model. (And the .png file made by snapshot in meshlab and the subsequent .jpg file are of correspondingly low resolution.)

There don't seem to be any parameters to set when exporting from agisoft, right?

Should I instead export to a different file type?  Which one works best?

Or to make it a more general question, what is the best way to get a high resolution jpg or tif from agisoft?

I'm using agisoft standard. 
If I buy professional, will the export results be higher resolution?

[James]

There shouldn't be any difference between the models, but one difference between photoscan and meshlab is that by default meshlab displays the model with artificial light shading, which will highlight any unsmoothness of your mesh, whereas photoscan does not do that, so you only ever see the colours as they are in the texture.

You can toggle this in meshlab by clicking the little yellow lightbulb in the middle of the toolbar at the top.

[Ethork]

Hmmm.  Maybe you're right, that its not the meshlab model that is less detailed - ie: the mesh, but rather the textures that are applied to it.   
The problem I'm having isn't due to "lighting on" vs. "lighting off"; the textured model in meshlab is considerably less detailed looking compared to the textured model in agisoft whether lighting is on or off. 

In any case, the textured model in meshlab looks simplified, the forms (irregular rocks in this case) are smoothed out and the junctions between them are smoothed out.  It may be a matter of meshlab not applying them to the model in sufficient detail.  Is there any way to tell it to apply more detail?   
(And this isn't a matter of "screen multiplier" in meshlab - I get the same lack of detail even at a high multiplier. (I've gone up to 18, anything above that and mesh lab crashes.))

What is the best way to output 2d images - jpgs or tifs - from agisoft?

[James]

I suggested meshlab because it seemed like the simplest way to get what you wanted, although it doesn't seem to be working out too well!

Pictures of the difference between the photoscan and meshlab model view might shed some light on the problem.

When i want high resolution images of my photoscan models i use Blender, but that's potentially rather more than a 5 point bullet list to get it working how you want! I'd say it was definitely worth looking into though. The things you need to do are import your model, remove any lights from the scene, set the material of your model to 'shadeless' in the blender internal render engine,  check that the texture is being applied (generally it just works automatically), set up a camera - i normally use orthographic rather than perspective but depends what you are doing, and hit f12 to render (You'll need to set render dimensions and output format too).

There are tons of tutorials out there but the trouble is that it's quite a lot of fun and you easily forget the serious reason you started doing it...

[stihl]

I'd suggest tilting the bounding box so the red side is parallel to the building facade.
Then creating a mesh heightfield and exporting this as an Orthomosaic from Photoscan.
At least that's what I did in the past, just let photoscan think that the building facade is the ground instead of a wall.

[AlecMoody]

You are probably noticing a difference in mesh normals. Photoscan uses all hard edges/no smoothing groups where meshlab is rendering smooth normals by default.