Forum

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.


Messages - David Cockey

Pages: 1 ... 3 4 [5] 6 7 ... 14
61
General / Re: EXR vs BMP panoramas
« on: July 14, 2014, 02:40:09 PM »
The difference in the appearance of the images is because the 32 bit EXR images do not have any gamma adjustment which is standard for images in 8 bit and 16 bit BMP, JPEG and TIFF formats.

Tonemapping the EXR images to a 8 bit or 16 bit format will make them look normal.

62
Feature Requests / Copy Camera command
« on: July 12, 2014, 10:06:55 PM »
A "Copy Cameras" command would be very useful for me. Same as "Move Cameras" except a copy of the cameras would remain in the source chunk.

63
Which operating system are you using?

131K points in the sparse cloud with 150 photos is reasonable.

A diagnostics run:

Copy the .psz file with a different name.

Start Task Manager and click on the Performance tab. How much memory is used without PhotoScan running?

Start PhotoScan and open the copy of the .psz file for the model. How much memory is used with PhotoScan open but not executing any commands?

Start to Build Dense Cloud.

Observe memory usage as shown in Task Manager. Let PhotoScan run until memory usage has leveled off. If it reaches 100% PhotoScan will start swapping data with the hard drive which makes PhotoScan run extremely slowly.

64
What settings did you use for Build Dense Cloud and Build Mesh?  60 hours with 150 photos sounds like the memory filled and swapped with the hard drive. If so one possible fix may be to increase memory size. Another fix is to split the project into smaller chunks after Align Photos.

65
General / Re: Plastic Fantastic Prme Lense for Canon
« on: July 10, 2014, 05:00:26 AM »
Yes, its the pancake lens. All the reviews and tests of it I've seen have been excellent and I've gotten very good results.

66
General / Re: Plastic Fantastic Prme Lense for Canon
« on: July 09, 2014, 11:41:21 PM »
I use the Canon EF 40mm f2.8 prime lens which has image quality very similar to the 50mm f1.8 with a wider field of view. It is a little more expensive, list price of $249 but typically available for $149 in the US. It has better build quality and quieter and probably faster autofocus than the 50mm f1.8 (which as Marcel points doesn't affect PhotoScan results).

67
General / Re: GPU processing during the modeling and settings
« on: July 08, 2014, 06:19:30 AM »
GPU is not used during Align Photos.

GPU is used heavily during Reconstructing depth portion of Build Dense Cloud.

GPU is not used during Build Mesh.

GPU is not used during Build Texture.

With I7-3770 CPU (4 cores / 8 virtual cores) and Radeon HD 7770 GPU best performance is with 6/8 "Active CPU Cores". However during Align Photos 100% of all 8 virtual cores are used.

68
Feature Requests / Duplicate photo removal
« on: July 06, 2014, 11:40:38 PM »
A command/feature for easy removal of photos with the same name from a chunk would be very helpful. I sometimes add photos to chunk twice by mistake.

69
General / Re: Alignment algorithm, does color matter?
« on: July 06, 2014, 11:37:53 PM »
My recollection is that Rhino works on color, not just brightness.

I've never had success using pieces of tape to add texture. Let us know how it works for you.

70
General / Re: Capturing a good dataset
« on: July 06, 2014, 06:44:07 PM »
Make sure the photos are properly exposed with no clipping on the glacier. If the photos  bright areas are over-exposed then those areas will not have any information PhotoScan can use. Any areas of sky in the photos can be overexposed. Use the camera's histogram to help set exposures. If possible set the histogram to show separate RGB curves rather than a single white composite curve.

71
Hi Morgan, another thing to look at is image stabilization. This should be switched off. Image stabilization tends to alter the focal length giving inconsistent results.

Image Stabilization doesn't alter the focal length (if it would do this, you would see the image wildly zooming when you look through the viewfinder with IS enabled).

But your advice is good anyway. When you are using a tripod Image Stabilization should indeed be turned off.  Otherwise the Image Stabilization would try to correct a non existing movement, and this actually adds blur to the photo.
Image Stabilization doesn't alter the focal length but it does shift the optical axis of the lens relative to the sensor. That's how it eliminates/minimizes blur due to motion. PhotoScan attempts to correct for any deviation of the lens optical axis from the center of the senor with the calibration parameters
cx, cy
Principal point coordinates, i.e. coordinates of lens optical axis interception with sensor plane.



72
General / Re: Is it possible to measure window openings with Ps?
« on: June 26, 2014, 12:37:54 AM »
Is the visual texture suitable for PhotoScan. Try a set of photos of several ground level windows and see if PhotoScan will process them with satisfactory results.

It sounds like the required level of accuracy is 2mm for each individual window opening, not an overall accuracy of 2mm. That can be a major difference.

73
Quote
David - I'm a little confused by your response. Do you mean that Photoscan may not model lens distortion properly?
PhotoScan models lens distortion properly, but it does not model lens distortion exactly for most lenses. Look close enough and the there will be a small amount of residual distortion for most lenses after PhotoScan corrects for distortion. Usually the amount of residual distortion is inconsequential  and doesn't affect the results for the intended purposes. The residual is likely to be smaller in magnitude than other sources of errors and uncertainties.

Quote
However, the lens distortion model in PhotoScan is pretty generic (and very well tested over multiple decades) and should be fine unless you have a particularly wacky lens (which I don't expect the D10 does).
The lens distortion model in PhotoScan should be fine. But the small differences between the actual lens distortion and the lens distortion model for each photo can add enough to be significant for some applications. With long, narrow objects where each photo used shows only a short section of the object I've had small magnitude, long wavelength distortions occur. Less overlap between photos tends to increase the amount of distortion.

74
General / Re: Manual photo alignment - tips?
« on: June 17, 2014, 05:42:47 PM »

Ah, so in Thomas case, his manually aligned cameras won't be selected because they don't have any points in the Sparse Cloud?

Is that all the Sparse Cloud does in relation to the Dense Cloud reconstruction (selecting which cameras are used)?
That is my understanding. A method is needed to generate the Sparse Cloud without re-aligning the cameras.

Also my understanding is Dense Cloud construction only uses the Sparse Cloud to select the camera pairs it uses.

75
General / Re: Manual photo alignment - tips?
« on: June 16, 2014, 07:39:45 PM »
Sparse points are very important to dense point cloud construction. Sparse points inside the reconstruction bounding box are used to select the camera pairs used for dense point cloud. This is why sometimes dense point cloud construction to fail if the box is too small.

Pages: 1 ... 3 4 [5] 6 7 ... 14