masking photos

Hi

I am using RC to survey buildings and to then extract colour ortho photos and mesh models.

The biggest problem I have is the work required to clean up the mesh and ortho photo’s where glazing in windows cannot be surveyed by the original scanner or identified by the software within RC.

It would be a great help if it were possible to create a feature which would allow us to mask the photo’s to indicate areas of glazing and to then have the software create a single mesh plane where the glass is so that the mesh model, texturing and ortho photo’s would then need much less work to clean them up.

Kind Regards

Robert

 

 

 

hello,

there have already been requests of this kind, please avoid creating duplicate posts in the same part of the forum,

I recommend you to use the search tool before starting a new topic,

following this rule helps us keep the requests better organised and managable,

thank you

I thought at first that it is a double post as well, But careful reading reveals that the request goes further then just masking. Robert suggests that the masks should be used to create a “forced” surface withing the geometry.

Personally, I find this suggestion extremely good, it would be very helpful - not just for us users but I imagine also for RC, since it wouldn’t be thrown off by those difficult areas.

 

+1 from me

+1 - sounds like a dream! Whole areas of rubbish flat featureless wall too. Beats creating planes in a modelling programme.

How about marking flat areas on one or more source images?

Draw a polygon around the area or something?

Its got my +1 :slight_smile:

I think that’s what Robert was aiming for, drawing a polygon on several images and creating a surface by doing so. Thinking about it, this is pretty much how the older kind of photogrammetry worked. The difference would be that some surfaces get integrated within the rest of the detailed model that RC creates…

Hopefully the DEVs will notice thisrequest after all.

That’s one way - another would be to draw the polygon in the model within RC.

However I realise that would require a higher level of 3D-modelling type tools within RC. Present rectangle and lasso tools are too imprecise, un-editable, and are vague in the view-z direction.

But still, would be great to be able to do it not to multiple photos but just once in the model within RC. Maybe such a special/limited-purpose 3D modelling tool wouldn’t be too difficult.

The advantage of doing it Control-Point-based in individual photos would be that the technology is already there. But that depends if it is possible at all to transfer info from there to the model which we have discussed elsewhere…  :slight_smile:

duplicate thread of: https://support.capturingreality.com/hc/en-us/community/posts/115001351072-image-masking