Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Topics - IanJohnson

Pages: [1]
General / Source Mesh for Reduce Overlap
« on: December 15, 2021, 06:46:40 PM »
How is the mesh used to calculate which images should be disabled for Reduce Overlap?  If I build a mesh from sparse cloud and reduce overlap, more photos are disabled than if I reduce overlap using a mesh built from low quality depth maps. 

I'm trying to figure out how much I can reduce overlap without getting holes.    Is the calculation more accurate using depth maps?  Does the quality of the depth maps matter?  If I reduce overlap, then build the model again with low quality depth maps, would that give me a heads up about where I will have holes in the final mesh?

Feature Requests / Chunk visibility
« on: July 02, 2011, 06:01:56 PM »
I would like a button to switch visibility of a chunk on and off.  I can see the chunk I am on, or all aligned, but it would be more useful to be able to see all aligned chunks and then switch chunks on and off. 

Also if you add manual alignment of chunks, an opacity control would be helpful too.

A completely unrelated request-  Would it be possible to allow the output view to be scrolled during processing?  Perhaps by breaking it into a separate window or writing to a log file?  Also timestamps with the events.

General / CPU Usage Drop
« on: June 30, 2011, 06:50:05 PM »

Often while generating a mesh on smooth, CPU usage will drop to 1-5%, and the process will never complete (maybe if I left it for days it would).  I thought it was caused by running out of memory, but I've seen it happen near the limit, but not exceeding.  Memory usage still varies, and there is an occasional spike, or one core might run at 20% for a little bit and drop off again.

This always seems to happen at 46% completion.  I dread 46%, and if it makes the jump to 82% I know the mesh will most likely complete.

It is currently doing it on a chunk with 69 10MP photos on a 3 core AMD with 8GB of RAM in Win7.  It's done the same thing on my work computer, with is a MacPro with 8 cores and 8GB RAM.

Feature Requests / Additional batch options
« on: June 27, 2011, 06:08:20 PM »

I would like to see Align Chunks and Merge Chunks added to the batch.  And in the Merge Chunks dialog an option to generate a texture for the new merged mesh.

I would also like an option to autosave after a long processes that are started outside of batch.

Bug Reports / Can't Cancel
« on: June 27, 2011, 05:06:28 PM »

For many processes, PS won't cancel if it has reached a challenging situation.  In particular, if it has run out of memory while building a mesh. It may not have completely run out of memory, but it is apparently running on cache, using 5% processor.   The "Are you sure?" dialog will come up and the Cancel button greys out, but PS keeps going.  You can even pause and resume, but it won't ever actually stop, leaving no option but to force quit. 

The same thing happens when aligning chunks and it gets stuck at 99% projecting points.

Feature Requests / Blend Vertex color with Photo Textures
« on: June 23, 2011, 07:26:20 AM »

Portions of the model that cannot be textured from a photo appear black.  It would be nice if those areas could be filled in and blended with the vertex color so the gaps aren't so obvious.

Feature Requests / Progressive Alignment of Chunks
« on: June 22, 2011, 08:55:31 PM »
I would like to be able to align a few chunks at a time and lock their orientation.    Then if I want to add and align additional chunks, I can reference them to one of the previously oriented chunks and leave the others disabled.  That way I don't have to run a time-consuming align of the entire project each time I want to add to it.

General / Chunk Alignment Question
« on: June 19, 2011, 08:09:36 PM »
Is chunk alignment based solely on the meshes, or does it include camera positions as well?  If one chunk is built at low quality, will it have trouble aligning with another chunk that is at medium?  Exact versus smooth?

I have a few chunks that are the side of a building and a few that are the front.  I'm having some trouble turning the corner.  I'm trying to build a single mesh at the corner that will bridge the front and side, but I'm having trouble getting it to align with any of the other chunks.   If I use some of the same photographs between chunks, can Photoscan use those common camera positions to assist in the alignment?

Feature Requests / Center Rotation on Object
« on: June 16, 2011, 08:15:57 PM »
When navigating a model, it would be helpful to have a button that will automatically center rotation on the center of the object.

Also hotkeys to switch between Rotate View, Select Faces, Bounding Box, and Rotate Bounding Box. 

Feature Requests / Batch Progress Report
« on: June 16, 2011, 02:45:35 AM »
When running a batch, I would like to see some indication of which chunk is processing, how many jobs were completed and how many remain.  Maybe something like a view of the batch window with each entry getting check off as it completes.

General / Out of Memory
« on: June 15, 2011, 06:18:34 AM »
I'm having trouble creating geometry from 13 5 megapixel photos on 4GB of RAM.  I ran a similar set with about 20 photos at High Quality 200k triangles and got a really nice model.  Other similar chunks keep failing. 

While reconstructing and creating geometry it runs at around 1.5GB of RAM.  When the geometry reaches 100%, memory usage begins to climb until it runs out and never begins decimating the mesh.  The same thing happens at 100k polys/High Quality, but a medium quality build will succeed.

Is this a bug, or is there something about a set of photos that can cause excessive memory usage between the mesh build and decimation?

This is with 0.8.2 build 979.  I started to try the same chunk with 0.8.1, but it was going to take 4 hours to complete the geometry rather than the 90 minutes or so that 0.8.2 was going to take and I didn't want to wait.  I'm trying 0.8.2 build 958 now.

Feature Requests / Automatic versioning
« on: June 15, 2011, 01:16:23 AM »

Photoscan crashed, or got force quit while shutting down, possibly while still saving my project file.  It seems that the project file got corrupted as a result, and won't open anymore.

It would be nice if while saving, Photoscan would keep previous saves up to a certain number so that if the project file was damaged, there would be a recent version to go back to.  I only lost today's processing because I have a nightly backup, but I would prefer to have lost only the last job.

General / Can't Merge Chunks
« on: June 14, 2011, 05:40:43 AM »
I have some chunks with lots of overlap, which aligned properly, but they won't merge- "Can't Merge Chunks".

The error message doesn't provide much guidance.  Why might the chunks not be able to merge, and how can I troubleshoot this? 

Also, can the new merged model be re-textured?  If I export it as .obj, do some editing in another application and then import it back into the same chunk, can I rebuild the texture?

Bug Reports / .obj imports with 0 faces
« on: June 13, 2011, 11:44:36 PM »
I downloaded the current version today.  I've found that any mesh I import as .obj results in an object with 0 faces.  This includes models that I've exported from Photoscan as .obj and tried to import back in, unchanged.  If I convert it to .ply it will work.

General / Aligning Point Clouds from Camera Positions
« on: May 06, 2011, 12:40:18 AM »
I've been using photoscan to align photos, then export the sparse cloud as bundle.out and using CMVS to create a dense cloud.  The advantage to this is once CMVS has split the job into parts I can divide the work between several computers.  This helps make up for the low amount of RAM I have (4GB) and distribute the processing.

For a large model of a building I thought it would help if I could run the alignment on a downsampled set of images to help run a lot of images without running out of memory.  Then use higher resolution images for the dense reconstruction in the areas that need more detail.

Using a test set of about 18 photos in a 180 degree view, I ran alignment on them at 10Megapixels, 5MP, and 3MP sizes.  I couldn't use the bundle.out from the 3MP set to reconstruct using the 10MP images.  The points were all wrong.  However if I used a 3MP bundle on a 3MP reconstruction, and 10MP bundle on a 10MP reconstruction, the two different point clouds lined up since they were based on the same images.

Suppose I align a large group of images covering a whole building facade, using images that I have downsampled to 3MP.  Then I take a smaller group of those images, which cover a decorated entrance, and align them using the 10MP versions.  The point cloud for that 10MP subset will probably be created at a different scale and orientation than the point cloud created from the whole set.

Would it be possible to add a feature which does this:  Create two chunks with a certain amount of images in common between them, and then base the scale and orientation of the point clouds on the camera calibrations derived from those shared images?

Also it would be nice if I could import a point cloud to be meshed within Photoscan, and have an opportunity to export the dense cloud created during the meshing process. 

I know there is alignment of meshes, and I haven't tried that yet.  I would like to be able to work on cleaning up the point clouds more before meshing, and having some kind of automatic alignment at that stage would be helpful.

Pages: [1]