About me Research Contact

Motion De-Blurring

If a camera moves fast while taking a picture, motion blur is induced. There exist techniques to prevent this effect to occur, such as moving the lens system or the CCD chip electro-mechanically. Another approach is to remove the motion blur after the images have been taken, using signal processing algorithms as post-processing techniques. For more than 30 years, numerous researchers have developed theories and algorithms for this purpose, which work quite well when applied to artificially blurred images. If one attempts to use those techniques to real world scenarios, they mostly fail miserably. In order to study why the known algorithms have problems to de-blur naturally blurred images we have built an experimental setup, which produces real blurred images with defined parameters in a controlled environment. On this page you will find a collection of de-blurring algorithms for you to study and motion blurred pictures in order to test your own algorithms.

Test Images

Our test images have been taken with a µC controlled camera, as shown in the picture on the left. The setup comprises a camera unit, a guiding rail and a stepper motor. The camera carriage is accelerated to a constant speed and takes a photo with a medium exposure time (around 100ms) to allow significant motion blur appear in the picture. As a starting motif a checkerboard structure has been chosen, since this allows an easy method for estimating the associated PSF. More details about the setup can be found in [1].


The following test images are available for deblur/download. They may be freely used for scientifc and non-commercial purpose. All other usage requires prior written permission by the owners.


File Size

Horizontal Displacement

File Size

Horizontal Displacement

Large a = 200px Large a = 200px
Medium a = 100px Medium a = 100px
Small a = 50px Small a = 50px


If you are interessed in studying the differences between real and synthetic motion blur, we have created synthetic versions of the upper left image. The images were created using a Matlab script also available.


Unblurred reference image for
comparision purposes

Blurred image with circular data
wrap around
File Size   File Size

Horizontal Displacement

Large Large a = 100px
Medium Medium a = 50px
Small Small a = 25px

Blurred image with repetetive data
wrap around

Blurred image with circular data
wrap around and added noise (AWGN, var = 0.01)
File Size

Horizontal Displacement

File Size

Horizontal Displacement

Large a = 100px Large a = 100px
Medium a = 50px Medium a = 50px
Small a = 25px Small a = 25px


So far, we know about the following de-blur algorithms. If you happen to know other algorithms we would be more than happy if you could email us.

Name Literature Matlab command Restoration Speed Restoration Quality
Direct Approach        
Wiener Filter   deconvwnr (IPT)    
Regularized Filer   deconvreg (IPT)    
Richardson-Lucy Deconvolution   deconvlucy (IPT)    
Maximum Likelihood Estimation   deconvblind (IPT)    
TU Berlin   deconv_tuberlin    
Sondhi   deconv_sondhi    
Advanced Landweber   deconv_alm    

Note: Some Matlab commands might require the Image Processing Toolbox (IPT) by Mathworks.


[1] Comparison of Motion Deblur Algorithms and Real World Deployment, S. Schuon,
K. Diepold, 2006, Paper on IAC 2006 (download as pdf)
[2] Nature of Motion Blur, S. Schuon (download as pdf)


  • Sebastian Schuon, schuon (at) mytum (dot) de
  • Klaus Diepold , kldi (at) tum (dot) de