Merging Chunks

[shroommu]

I'm a very new user of Photoscan, and I've run into a question on merging chunks for which I haven't seen an answer in the handbook or in the forum.

I'm working on a model of a riverbed. Half of the riverbed is in shadow, and the other half is well lit. I couldn't get both sides to show up in one chunk, so I have created a chunk for each side. Now I'm trying to merge the chunks, but I'm not getting the result that I expected.

I have the standard version of Photoscan, so the workflows I've tried so far are:

Align Chunks>Point Based>Merge Models
Align Chunks>Camera Based>Merge Models

What I expect should happen is that Photoscan will create a merged chunk using the two chunks, and I'll then end up with a single mesh which includes both the light and dark side of the riverbed. The actual result I get is a merged chunk that is just a duplicate of one of the other chunks.

What am I doing wrong?

[Alexey Pasumansky]

Hello shroommu,

Chunks should be aligned prior to the merging - you should see [R] mark next to their labels in the Workspace pane.

Camera based approach wouldn't work unless you have the images with the same labels in both chunks that represent exactly the same camera positions in the real world.
Point based approach means that PhotoScan will be trying to find the matching points for the images from the different chunks, so if the image matching operation fails completely for the complete image set and no tie points are found for the images from light and dark sub-sets, the point based chunk alignment would likely fail.

[Alexx]

Hi,
I have a similar task to solve: I have exact the same chunk twice, but with different parts deleted from the mesh ("3D model").
All I need is to combine them.
I aligned them (by cameras, because they were the same) and the result is indeed the 3D model I want to have.
BUT all cameras are doubled and the size of the dense cloud is exactly twice as big.

It means that Agisoft is not smart enough to understand, that the cameras are the same and the dense clouds are identical. How can I improve that? How can I automatically remove duplicate cameras and duplicate dense cloud points?

And how in general can I move the 3D model between chunks (that have identical cameras and dense cloud)?

Do I have to export (obj) and import again? I tried and it didn't work: although I have the correct size of the 3Dmodel (bigger number of faces), the new added faces are not visible (still deleted)

Thanks