Messages

136

Bug Reports / Re: photoscan 1.0 and 1.0.1 unable to correctly open file generated with 0.9 on linu

« on: February 07, 2014, 11:26:16 PM »

I have the same problem with files created in 0.9. I'm currently running 1.0.1 Build 1812 64 bit with Windows 8.1

137

Face and Body Scanning / Re: Advantages of using full frame?

« on: February 07, 2014, 08:58:37 PM »

Resolution depends on the sensor in the camera. Some full frame cameras have less resolution than some crop frame cameras.

138

Face and Body Scanning / Re: Automatic Masking

« on: February 05, 2014, 07:14:02 AM »

Coded targets are not needed to create a mask from the background.

139

General / Re: Automatic mask in PhotoShop

« on: February 05, 2014, 07:12:09 AM »

I did not know that I do it also with PhotoScan. thanks !

another question, do you know how the option "mask by a model" works? What model exactly it refers and in which cases is more handy ?

thanks a lot 

For mask by model to work, you need to reconstruct 3d model, export it (can edit in external app) and use it as mask. With new build dense meshes option, for me it is better option then mask by a model.

If you have static cameras (turntable for example) i used to make background image by hand in Photoshop ,painting similar colors over subject. and it worked well. It is usefull if you forget to take background image or it is faulty for some reason.

No need to export the 3D model to use it in "mask by model". Build the mesh in PhotoScan and trim as desired. Then right click on a photo, select Import Mask, select "From Model" as method, select what masks should be applied in "Import masks for", and click okay. Masks will be created based on the mesh. Areas outside of the model will be masked in the photos.

I frequently use Mask from Model in situations where the background is moving relative to the object by a small amount, such as a boat tied to a dock or an object which was bumped while being photographed. I align photos, create a dense point cloud and build an initial mesh without masking. Masks are then created using the initial mesh. The masked photos are re-aligned. The masks are then removed and a new dense point cloud is created and mesh built.

140

General / Re: Cannot export DEM model, please help

« on: February 04, 2014, 05:40:30 AM »

Which version and release of PhotoScan? Which format are you trying to export in?

141

General / Re: Rough surface problem...

« on: January 21, 2014, 06:24:08 AM »

My understanding is most/all Nikon DSLR's have ISO 200 as the base ISO and there is no reduction in noise with ISO 100.

Which "quality" settings are you using for alignment and dense cloud generation?

142

General / Re: Ultra High has more occlusions and holes than High

« on: January 16, 2014, 04:56:47 AM »

I've had the same happen. It appears to depend on the "scale" of the visual texture as it appears relative to the size of the pixels in the photos and the corresponding differences in color/brightness between adjacent pixels. Holes appear to occur where the scale of the visual texture is multiple pixel widths.

Also, blurry photos cause problems at higher quality levels.

143

General / Re: Long Processing time

« on: January 16, 2014, 04:48:22 AM »

Hello ricky4207,

Please check in your system monitor that the process has not gone to swap. If there is not sufficient memory you'll need to lower the quality.

Or sub-divide the model after aligning the photos into several chunks.

I've been using Open Hardware Monitor shareware to monitor system usage including memory with Windows. Not perfect but it does record maximum usage since last reset.

How long does Dense Cloud generation with Low and Medium quality take?

Also, check that GPU usage is enabled in Preference - OpenGL. The appropriate box needs to be checked.

What graphics card / GPU do you have?

144

General / Re: Computer requirements...

« on: January 15, 2014, 05:08:17 PM »

A superfast computer with very large amounts of memory is not absolutely required for processing relatively large projects.

I routinely process projects with 200 to 750 15MP photos on a computer with:
CPU - Intel I7 3770 Quad Core
GPU - AMD 3770, 3 GB memory (PhotoScan only uses 2GB)
Memory - 16MB, DDR3-1600
OS - Windows 7

Projects are "arbitrary" geometry, high accuracy is used for alignment, medium or high resolution for dense cloud generation with ultrahigh for small areas as needed.

Typical dense cloud generation performance:
Device 1 performance: 111.838 million samples/sec (CPU)
Device 2 performance: 328.192 million samples/sec (Capeverde)
Total performance: 440.03 million samples/sec

I divide the project into chunks as needed to stay within the memory limitation. Each stage may have a different number of chunks. For dense point cloud generation the bounding box is carefully set, and with large numbers of photos the object is divided into several overlapping chunks.

145

General / Re: Align photos point limit

« on: January 08, 2014, 07:08:09 PM »

The point limit sets the maximum number of points which will be identified in each photo during the Align Photos process. As cbnewham mentioned the number of points actually identified on each photo may be less than the maximum number set.

I have experimented setting the maximum number to 80,000 with High Accuracy when aligning 15Mp photos. I have not seen appreciable difference compared to a limit of 40,000.

With Medium Accuracy only 20,000 points or less are identified in each photo so changing from 40,000 to 80,000 does not make any difference.

 

146

General / Re: CPU and GPU benchmarks

« on: January 08, 2014, 06:25:33 PM »

.....

By recommendation from Agisoft, the GPU numbers have two CPU cores disabled each.  Meaning that Stage 1/3/4, which do not use the GPU, might look slower than when the GPU is not used.

.....

I have two computers, each with a quad core I7 and a GPU. PhotoScan on each is set with "2 CPUs disabled". During much of the Align Photos and Build Mesh all CPUs are running at or very close to 100% according to Windows Performance Monitor and Open Hardware Monitor. There is no difference in performance during these stages when the number of CPUs "disabled" in PhotoScan is changed.

147

General / Re: 8 vs 16 bit Images

« on: January 06, 2014, 02:07:02 AM »

A 32 bit pseudo "HDR" created from one RAW file does not contain any additional information compared to a 16 bit TIFF created from the same one RAW file. RAW files have 12 or 14 bits of precision.

32 bit pseudo "HDR"s from a single RAW files may be worse for PhotoScan if the pseudo "HDR" contains artificial details which are inconsistent from photo to photo.

A HDR created from multiple RAW files with properly bracketed exposure may have additional information if the dynamic range of the scene was large enough that the dynamic range of the camera was exceeded. In that case the HDR creation combines data from the multiple RAW files.

If the dynamic range of the scene fits within the dynamic range of the camera then an HDR from multiple RAW files will not contain any additional information compared to a 16 bit TIFF provided proper exposure was used.

148

General / Re: 8 vs 16 bit Images

« on: January 06, 2014, 01:41:29 AM »

Version 1.0.0 Build 1795 is now the standard release.

149

General / Re: Interior scan

« on: January 05, 2014, 06:25:53 AM »

Using "32 bit pseudo hdr from one raw photo" is no better then using 16 bit TIFF from RAW. Depending on the how the "32 pseudo hdr" is generated the result may even be worse for PhotoScan.

Your Canon 550D captures 12 bits per pixel in RAW, and processing the photos to generate 32 bit photos does not add any information. Use 16 bit TIFFs which will contain the 12 bit data available in the RAWs, and keep the EXIF information. I have a Canon 500D and use 16 bit TIFFs as my standard input into PhotoScan. I have converted RAW to 16 bit TIFF for PhotoScan using Canon Digital Photo Professional (included on the disk which came with the camera), Lightroom and OpticsPro.

A 30 mm focal length lens on a Canon 550D is equivalent to a 48 mm lens on a full frame 35 mm camera. Sensor size of 550D is 22.3mm x 14.9mm which results in a crop factor of 1.61 = 36/22.3 = 24/14.9. 1.524 is the crop factor of a Nikon DX.

150

General / Re: Agisoft PhotoScan 1.0.0 pre-release

« on: December 22, 2013, 08:30:46 AM »

PhotoScan Version 1.0.0 build 1789 (64 bit) Windows 8.1 crashes and PhotoScan closes immediately when Align Chunks using Point Based method is used.

Align Chunks using Camera Based method works.

Version 9.1 build 1714 works with same files (saved as version 9.1 type) for both Align Chunks methods.


Crash reporter contents:

ARB_texture_non_power_of_two: Yes
ARB_vertex_buffer_object: Yes
OpenGLMaxTextureSize: 16384
OpenGLRenderer: AMD Radeon HD 7700 Series
OpenGLStereo: No
OpenGLVendor: ATI Technologies Inc.
OpenGLVersion: 4.2.12422 Compatibility Profile Context 13.152.0.0
ProductName: PhotoScan Standard Edition
System: Windows 64bit
SystemMemory: 16.0 GB
Throttleable: 1
URL:
Vendor: Agisoft
Version: 1.0.0.1789
XLog: OpenGL Vendor: ATI Technologies Inc.
OpenGL Renderer: AMD Radeon HD 7700 Series
OpenGL Version: 4.2.12422 Compatibility Profile Context 13.152.0.0
Maximum Texture Size: 16384
Quad Buffered Stereo: not enabled
ARB_vertex_buffer_object: supported
ARB_texture_non_power_of_two: supported
Loading project...
loaded project in 0.078 sec
Finished processing in 0.078 sec (exit code 1)


This report also contains technical information about the state of the application when it crashed.