Depth from Focus

From MMVLWiki
(Difference between revisions)
Jump to: navigation, search
m (→3D surface metrology)
m
 
(35 intermediate revisions by one user not shown)
Line 1: Line 1:
−
{|align="right"
 
−
|-
 
−
|[[Image:Dvcoin.png|thumb|240px|Extended depth-of-field image showing part of a 10-pence coin (compare with 1.39 MByte [http://vision.eng.shu.ac.uk/jan/coin.avi video of focus stack])]]||[[Image:Coin.png|thumb|260px|3D reconstruction (profile is exagerated)]]
 
−
|-
 
−
|}
 
 
=Depth from Focus=
 
=Depth from Focus=
 
==3D surface metrology==
 
==3D surface metrology==
−
* In principle it should work with any microscope, which has a motorized z-drive and a digital camera (see [[Mimas Video Input|Mimas video input]]).
+
* You need to grab the focus stack with your microscope and digital camera.
 
* The results will be even better, if the illumination optics of the microscope can project a pattern.
 
* The results will be even better, if the illumination optics of the microscope can project a pattern.
 
* Non-destructive measurement of surface profiles
 
* Non-destructive measurement of surface profiles
 
* With our experimental settings we observed (of course the result will depend on the quality of the microscope)
 
* With our experimental settings we observed (of course the result will depend on the quality of the microscope)
−
** Vertical resolution up to <math>0.2\ \mu m</math> (depending on aperture-size, magnification, projection-pattern and the surface properties of the object).
+
** Vertical resolution <math>\ge 0.2\ \mu m</math> (depending on aperture-size, magnification, projection-pattern and the surface properties of the object).
−
** Lateral resolution up to <math>2\ \mu m</math> (depends).
+
** Lateral resolution <math>\ge 2\ \mu m</math> (depends).
 
* '''Open Source''' (you are free to improve the code yourself if you redistribute it).
 
* '''Open Source''' (you are free to improve the code yourself if you redistribute it).
  
 
In general one can say: ''The lower the depth of field, the higher the resolution of the reconstruction.''
 
In general one can say: ''The lower the depth of field, the higher the resolution of the reconstruction.''
−
With high magnification (assuming constant numerical aperture) the resolution of the reconstruction goes up. The trade-off is that the reconstruction will cover a smaller area. This could be overcome by lateral stitching (e.g. cogwheel below).
+
With high magnification (assuming constant numerical aperture) the resolution of the reconstruction goes up. The trade-off is that the reconstruction will cover a smaller area. This can be overcome by lateral stitching (e.g. cogwheel below).
  
 
==Demonstration==
 
==Demonstration==
−
Here are some typical microscope images (showing a surface, which has been [http://www.twi.co.uk/j32k/unprotected/band_1/surfi-sculpt_index.html shaped using a power beam]).
+
Here are some typical microscope images (Surfi-Sculpt refers to objects, which has been [http://www.twi.co.uk/j32k/unprotected/band_1/surfi-sculpt_index.html shaped using a power beam]).
 
{|align="center"
 
{|align="center"
 
|-
 
|-
−
|[[Image:grid_0055.jpg|thumb|200px|First surfi-sculpt object, Leica DM LAM]]||[[Image:wheel_0136.jpg|thumb|200px|Second surfi-sculpt object, Leica DM LAM]]||[[Image:BrokenSuevit1_001500.png|thumb|200px|Piece of Suevit (enamel like material from meteorite impact) from the [http://en.wikipedia.org/wiki/Ries Nördlinger Ries], Leica DM RXA]]||[[Image:Hair20.png|thumb|200px|Micro-camera image of a hair on top of a laser printout]]
+
|[[Image:grid_0055.jpg|thumb|180px|First surfi-sculpt object, Leica DM LAM]]||[[Image:wheel_0136.jpg|thumb|180px|Second surfi-sculpt object, Leica DM LAM]]||[[Image:BrokenSuevit1_001500.png|thumb|180px|Piece of Suevit (enamel like material from meteorite impact) from the [http://en.wikipedia.org/wiki/Ries Nördlinger Ries], Leica DM RXA]]||[[Image:Hair20.png|thumb|180px|Micro-camera image of a hair on top of a laser printout]]||[[Image:Coin063.jpg|thumb|180px|Surface of 10-pence coin, Leica DM LAM]]
 
|-
 
|-
 
|}
 
|}
Line 28: Line 23:
 
{|align="center"
 
{|align="center"
 
|-
 
|-
−
|[[Image:griddv.jpg|thumb|200px|Extended depth of view for first object]]||[[Image:wheeldv.jpg|thumb|200px|Extended depth of view for second object]]||[[Image:SuevitDV.png|thumb|200px|Extended depth of view for Suevit (fringes have been removed manually)]]||[[Image:HairDV.png|thumb|200px|Extended depth of view for the hair]]
+
|[[Image:griddv.jpg|thumb|180px|Extended depth of view for first object]]||[[Image:wheeldv.jpg|thumb|180px|Extended depth of view for second object]]||[[Image:SuevitDV.png|thumb|180px|Extended depth of view for Suevit (fringes have been removed manually)]]||[[Image:HairDV.png|thumb|180px|Extended depth of view for the hair]]||[[Image:Dvcoin.png|thumb|180px|Extended depth-of-field image showing part of the 10-pence coin (compare with 1.39 MByte [http://vision.eng.shu.ac.uk/jan/coin.avi video of focus stack])]]
 
|-
 
|-
 
|}
 
|}
Line 35: Line 30:
 
{|align="center"
 
{|align="center"
 
|-
 
|-
−
|[[Image:grid1080.jpg|thumb|200px|3D reconstruction of first object [http://vision.eng.shu.ac.uk/jan/grid1.avi (742kB video)]]]||[[Image:wheel1008.jpg|thumb|200px|3D reconstruction of second object [http://vision.eng.shu.ac.uk/jan/wheel1.avi (725kB video)]]]||[[Image:suevit20.png||thumb|200px|3D reconstruction of suevit [http://vision.eng.shu.ac.uk/jan/suevit.mpg (1.4MB video)]]]||[[Image:Hair.png|thumb|200px|3D reconstruction of hair [http://vision.eng.shu.ac.uk/jan/hair.mpg (1.4MB video)]]]
+
|[[Image:grid1080.jpg|thumb|180px|3D reconstruction of first object [http://vision.eng.shu.ac.uk/jan/grid1.avi (742kB video)]]]||[[Image:wheel1008.jpg|thumb|180px|3D reconstruction of second object [http://vision.eng.shu.ac.uk/jan/wheel1.avi (725kB video)]]]||[[Image:suevit20.png||thumb|180px|3D reconstruction of suevit [http://vision.eng.shu.ac.uk/jan/suevit.mpg (1.4MB video)]]]||[[Image:Hair.png|thumb|180px|3D reconstruction of hair [http://vision.eng.shu.ac.uk/jan/hair.mpg (1.4MB video)]]]||[[Image:Coin.png|thumb|180px|3D reconstruction of 10-pence coin surface. The profile is amplified 5 times [http://vision.eng.shu.ac.uk/jan/coin3D.avi (1.0 MByte video)]]]
 
|-
 
|-
 
|}
 
|}
Line 41: Line 36:
 
{|align="center"
 
{|align="center"
 
|-
 
|-
−
|[[Image:Grid245.jpg|thumb|200px|Red-cyan anaglyph image of first object [http://vision.eng.shu.ac.uk/jan/grid.avi (2.0MB video)]]]||[[Image:Sugar.jpg|thumb|200px|Extended depth of field image for a piece of sugar (compare with 532 kByte [http://vision.eng.shu.ac.uk/jan/sugarstack.avi video of focus stack])]]||[[Image:Fiberdf10xdv.png|thumb|200px|Extended depth of field image of glass fibers (compare with 1.09 MByte [http://vision.eng.shu.ac.uk/jan/fiberdf10x.avi video of focus stack])]]||[[Image:SmallWheel.png|thumb|200px|Reconstruction of ''0.6 mm'' cogwheel using multiple focus sets]]
+
|[[Image:Grid245.jpg|thumb|180px|Red-cyan anaglyph image of first object [http://vision.eng.shu.ac.uk/jan/grid.avi (2.0MB video)]]]||[[Image:Sugar.jpg|thumb|180px|Extended depth of field image for a piece of sugar (compare with 532 kByte [http://vision.eng.shu.ac.uk/jan/sugarstack.avi video of focus stack])]]||[[Image:Fiberdf10xdv.png|thumb|180px|Extended depth of field image of glass fibers (compare with 1.09 MByte [http://vision.eng.shu.ac.uk/jan/fiberdf10x.avi video of focus stack])]]||[[Image:SmallWheel.png|thumb|180px|Reconstruction of ''0.6 mm'' cogwheel using multiple focus sets]]||[[Image:Icing.png|thumb|180px|Icing of a cake as seen under the microscope]]
 +
|-
 +
|[[Image:dvtungsten3.png|thumb|180px|Optical microscope image of tungsten tip]]||[[Image:dvtungsten6.png|thumb|180px|Optical image of tungsten tip (high magnification)]]||||||
 
|-
 
|-
 
|}
 
|}
  
−
As the idea for the algorithm was fixed already, it was possible to implement the algorithm as a command-line tool in less than 4 days, using existing [[Mimas]]-software (exspecially the [http://vision.eng.shu.ac.uk/mimas-docs/arrayOperators.html operators for boost::multi_array]).
+
As the idea for the algorithm was fixed already, it was possible to implement the algorithm as a command-line tool in less than 4 days, using existing [[Mimas]]-software (exspecially the [http://vision.eng.shu.ac.uk/jan/mimas/docs/arrayOperators.html operators for boost::multi_array]).
  
 
As this is a "quick hack", there's still lots of space for improvements.
 
As this is a "quick hack", there's still lots of space for improvements.
  
 
==Download==
 
==Download==
−
[[Image:New.gif|New.gif]] The software for estimating height-maps and images with extended depth-of-field is available for free (under the LGPL)! You first need to install [[Mimas#Mimas-2.0|version 2.0]] of the [[Mimas|Mimas Real-Time Computer Vision Library]] to be able to compile and run [http://vision.eng.shu.ac.uk/jan/depthoffocus-0.1.tar.bz2 depthoffocus-0.1] (652 kByte). The software also comes with sample files to generate photo-realistic 3D-reconstruction using [http://www.povray.org/ POVRay]!
+
===Old C++ Implementation===
 +
The software for estimating height-maps and images with extended depth-of-field is available for free (under the LGPL)! You first need to install [[Mimas_Announcement_Archive#Mimas-2.0|version 2.0]] of the [[Image:Mimasanim.gif|40px]] [[Mimas|Mimas Real-Time Computer Vision Library]] to be able to compile and run [http://vision.eng.shu.ac.uk/jan/depthoffocus-0.1.tar.bz2 depthoffocus-0.1] (652 kByte). The software also comes with sample files to generate photo-realistic 3D-reconstruction using [http://www.povray.org/ POVRay]!
 +
===New Ruby Implementation===
 +
[[Image:New.gif]] I've ported the program to [http://www.ruby-lang.org/ Ruby]. The new implementation is based on the [[Image:Hornetseye.png|40px]] [[HornetsEye]] Ruby-extension. It also requires [http://trollop.rubyforge.org/ TrollOp]. You can copy-and-paste the source code from the '''[http://www.wedesoft.demon.co.uk/hornetseye-api/files/stack-txt.html depth from focus example]''' page. The source code also comes with the [[HornetsEye]] source package.
 +
 
 +
=See Also=
 +
* [[MiCRoN]]
 +
 
 +
=External Links=
 +
* Publications
 +
** M. Boissenin, J. Wedekind, A.N. Selvan, B.P. Amavasai, F. Caparrelli, J.R. Travis: [http://dx.doi.org/10.1016/j.imavis.2006.03.009 Computer vision methods for optical microscopes]
 +
** J. Wedekind: [http://vision.eng.shu.ac.uk/jan/MechRob-paper.pdf Focus set based reconstruction of micro-objects]
 +
* Further Reading
 +
** [http://en.wikipedia.org/wiki/Anaglyph_image Wikipedia article on anaglyph images]
 +
** Ishita De, Bhabatosh Chanda, Buddhajyoti Chattopadhyay: [http://dx.doi.org/10.1016/j.imavis.2006.04.005 Enhancing effective depth-of-field by image fusion using mathematical morphology]
 +
* Focus stitching software [http://www.hadleyweb.pwp.blueyonder.co.uk/ CombineZ5] and some [http://www.janrik.net/insects/ExtendedDOF/LepSocNewsFinal/EDOF_NewsLepSoc_2005summer.htm impressive examples by Rik Littlefield].
 +
* [http://hirise.lpl.arizona.edu/anaglyph/index.php NASA anaglyph images of Mars]
 +
* [http://www.youtube.com/results?search_query=yt3d%3Aenable%3Dtrue 3D videos on Youtube]
 +
* [http://www.heliconsoft.com/ Helicon] proprietary focus stitching software
 +
 
 +
{{Addthis}}
  
−
=Links=
+
[[Category:Projects]]
−
* [http://en.wikipedia.org/wiki/Anaglyph_image Anaglyph images]
+
[[Category:Micron]]
−
* [http://www.ubka.uni-karlsruhe.de/cgi-bin/psview?document=2002/informatik/2 german diploma thesis] about depth of focus (also available as [http://www.ubka.uni-karlsruhe.de/vvv/2002/informatik/2/2.pdf PDF])
+
−
* Mechrob-paper ''[http://www.shu.ac.uk/mmvl/research/MechRob04-paper.pdf Focus set based reconstruction of micro-objects]''
+
−
* [http://www.mrt.uni-karlsruhe.de/hz_prj1.php Depth-from-focus applied in forensic science, MRT Karlsruhe]
+
−
* Focus stitching software [http://www.hadleyweb.pwp.blueyonder.co.uk/ CombineZ5] and examples at [http://www.micropics.org.uk/ http://www.micropics.org.uk/]
+
−
[[Category:Mimas]]
+

Latest revision as of 18:00, 17 March 2010

Contents

[edit] Depth from Focus

[edit] 3D surface metrology

  • You need to grab the focus stack with your microscope and digital camera.
  • The results will be even better, if the illumination optics of the microscope can project a pattern.
  • Non-destructive measurement of surface profiles
  • With our experimental settings we observed (of course the result will depend on the quality of the microscope)
    • Vertical resolution <math>\ge 0.2\ \mu m</math> (depending on aperture-size, magnification, projection-pattern and the surface properties of the object).
    • Lateral resolution <math>\ge 2\ \mu m</math> (depends).
  • Open Source (you are free to improve the code yourself if you redistribute it).

In general one can say: The lower the depth of field, the higher the resolution of the reconstruction. With high magnification (assuming constant numerical aperture) the resolution of the reconstruction goes up. The trade-off is that the reconstruction will cover a smaller area. This can be overcome by lateral stitching (e.g. cogwheel below).

[edit] Demonstration

Here are some typical microscope images (Surfi-Sculpt refers to objects, which has been shaped using a power beam).

First surfi-sculpt object, Leica DM LAM
Second surfi-sculpt object, Leica DM LAM
Piece of Suevit (enamel like material from meteorite impact) from the Nördlinger Ries, Leica DM RXA
Micro-camera image of a hair on top of a laser printout
Surface of 10-pence coin, Leica DM LAM

Using a focus-stack one can compute images with extended depth of focus:

Extended depth of view for first object
Extended depth of view for second object
Extended depth of view for Suevit (fringes have been removed manually)
Extended depth of view for the hair
Extended depth-of-field image showing part of the 10-pence coin (compare with 1.39 MByte video of focus stack)

If the surface can be illuminated properly, one can even do a 3D-reconstruction of the surface:

3D reconstruction of first object (742kB video)
3D reconstruction of second object (725kB video)
 [[Image:suevit20.png 180px|3D reconstruction of suevit (1.4MB video)]]
3D reconstruction of hair (1.4MB video)
3D reconstruction of 10-pence coin surface. The profile is amplified 5 times (1.0 MByte video)
Red-cyan anaglyph image of first object (2.0MB video)
Extended depth of field image for a piece of sugar (compare with 532 kByte video of focus stack)
Extended depth of field image of glass fibers (compare with 1.09 MByte video of focus stack)
Reconstruction of 0.6 mm cogwheel using multiple focus sets
Icing of a cake as seen under the microscope
Optical microscope image of tungsten tip
Optical image of tungsten tip (high magnification)

As the idea for the algorithm was fixed already, it was possible to implement the algorithm as a command-line tool in less than 4 days, using existing Mimas-software (exspecially the operators for boost::multi_array).

As this is a "quick hack", there's still lots of space for improvements.

[edit] Download

[edit] Old C++ Implementation

The software for estimating height-maps and images with extended depth-of-field is available for free (under the LGPL)! You first need to install version 2.0 of the Mimasanim.gif Mimas Real-Time Computer Vision Library to be able to compile and run depthoffocus-0.1 (652 kByte). The software also comes with sample files to generate photo-realistic 3D-reconstruction using POVRay!

[edit] New Ruby Implementation

New.gif I've ported the program to Ruby. The new implementation is based on the Hornetseye.png HornetsEye Ruby-extension. It also requires TrollOp. You can copy-and-paste the source code from the depth from focus example page. The source code also comes with the HornetsEye source package.

[edit] See Also

[edit] External Links

Bookmark and Share

Retrieved from "https://vision.eng.shu.ac.uk/mmvl/mmvlwiki/index.php?title=Depth_from_Focus&oldid=5523"
Personal tools
Namespaces
Variants
Actions
Navigation
Toolbox