Messages

46

General / Re: Can't build Dense Point Cloud

« on: April 10, 2015, 10:29:02 AM »

Glad to hear you found the problem.

I would set the Key Points to 40.000 and the Tie Points to 0. Your alignment will take a bit longer (but not excessively long), but it will be more accurate. Setting Pre-selection to "Generic" speeds up the alignment and generally has no downsides.

47

General / Re: Can't build Dense Point Cloud

« on: April 09, 2015, 10:57:28 AM »

How many points are in your Sparse Cloud, and what Tie Point setting did you use during the alignment (Tie Points are under Advance Settings in the alignment window)?

48

General / Re: Multiple gpus

« on: April 08, 2015, 07:58:29 PM »

SLI is not needed for Photoscan (SLI is more for rendering realtime computer graphics, Photoscan is using OpenCL).

I have disabled all my cores and let the GPU do all the work. Compared to the GPU, the performance of the CPU is really low, especially for your dual Titan setup.

With all CPU cores disabled you can also run multiple scans at the same time without much slowdown. When one scan enters the Depth Map Filtering phase (which is CPU only) you can start the Dense Cloud build for a second scan.

49

General / Re: Larger battery for EOS 500D

« on: April 08, 2015, 04:40:45 PM »

The biggest power user of the camera is often the LCD screen. So not using Live View and disabling the photo review (showing the photo on your screen after the shot) helps a lot with making your battery last longer.

50

General / Re: Field explanation of photoscan .xml camera file

« on: April 07, 2015, 08:34:26 PM »

Can't you export the scene to FBX? I think that includes the cameras as well (I don't have Maya or Max so I cannot check).

51

General / Re: Field explanation of photoscan .xml camera file

« on: April 07, 2015, 02:32:35 PM »

I think this is for the region box? (The box you define to tell where Photoscan will build the Dense Cloud and Mesh.)

52

General / Re: Photography Tip!

« on: April 07, 2015, 02:30:32 PM »

I'm not sure if that is possible without a multi camera rig.  With a multi camera rig you could Align and build the Dense Cloud with photos of your object in 'covered' state, and then build the texture in 'clean' state. (the photos are in the exact same position). If you are using handheld photos or a turntable then it will be very hard.

53

General / Re: What does Optimize Cameras actually do?

« on: April 07, 2015, 10:28:53 AM »

As far as I know, Optimize Cameras re-aligns the cameras based on the points in the Sparse Cloud*.

This is useful in combination with Gradual Selection: first you remove inaccurate points from the Sparse Cloud and then you re-align the cameras. The result should be a more accurate alignment of the cameras.

If you have a good alignment to begin with, you will probably not see a change in the positions or angles of the cameras. This is because a good alignment already has a very high accuracy (the same point viewed through two different cameras might only be offset by a single pixel, or even less). But for Depth Reconstruction at High or Ultra quality it might still be useful to Optimize the alignment (if your alignment is inaccurate it can lead to noise in the Dense Cloud).

* it probably also does some other stuff like optimize lens distortion correction. Compared to the original alignment it also takes much less time (not counting 'Extracting Points' and 'Comparing Points'), which puzzles me a bit.

54

General / Re: Photography Tip!

« on: April 07, 2015, 10:15:56 AM »

If you need to downscale your photos t get good results, it generally means your photos are not sharp.

Could you show a few source photos at full resolution? It's hard to give advice without seeing exactly what you do.

55

General / Re: Best practices for tank scanning

« on: March 29, 2015, 09:26:42 PM »

tarkhilIf still can go to the tanks can guide how precisely can capture them.. contact me over PM...

Why not post some tips here, that way everybody benefits?

56

General / Re: Alignment Experiments

« on: March 29, 2015, 11:52:12 AM »

You can set the Key point Limit to 0 for maximum points as well (no need to set it to 1 million)

5) Repeat steps 3-4 until the reconstruction error is minimized

Does repeated optimization lower the reconstruction error a lot? I have never tried that, thanks for the tip!

57

General / Re: Photoscan giving me (two) different alignment point clouds (Bug?)

« on: March 28, 2015, 11:00:17 AM »

However what I did notice, even though achieving a much superior result is that my generated model is still quite noisy, my photos are well exposed and shot at the lowest ISO (160) and even the noise within the photos was cleaned up, I don't know what could be causing a bumpy mesh

A bumpy / noisy mesh could be because of blurry photos. Go through your photos and see if they are in focus at the pixel level. Remove photos that are completely out of focus or show camera shake, and re-align the whole set or use Tools->Optimize Cameras. Be very strict about photo quality.

Another issue can be when the shape of the object changes in photos (ie. your subject moved their head or changed expression). It's hard to keep perfectly still during the whole shoot.

If you use too much noise reduction you might remove the pixel detail that Photoscan uses for reconstruction, so be careful not to crank the noise reduction too high.

Even the super 100+ camera rigs can show some noise in harder to scan areas, so don't expect perfect output.

58

General / Re: Best practices for tank scanning

« on: March 27, 2015, 07:08:53 PM »

I still don't see how having 5 images with 80% overlap helps. Please see the attachment of what I mean with 5 shots vs 2 shots.

Especially when using a panorama head, the 3 extra images will have the very same information. There is no extra useful information for Photoscan to reconstruct anything?

59

Feature Requests / "Show/Hide Cameras" shortcut key

« on: March 27, 2015, 03:24:20 PM »

I use Show/Hide Cameras a lot, because often the cameras obstruct the scan/region.

It would be really nice if there was a shortcut key for this (for example SHIFT-5 to toggle).

60

General / Re: Best practices for tank scanning

« on: March 27, 2015, 02:28:43 PM »

But I don't think it helps taking 5 shots that have a lot of overlap.

Unless you use a pano rig and group the images as a camera station...

I haven't used the camera station function myself, but I assume it treats all shots in the camera station as a single panoramic photo?

I can't see how it would be useful to have 5 shots at the same position with 80% overlap. You are not adding more information compared to 2 shots with 20% overlap. (even more so when using a pano rig where there is no parallax at all).