Sobel operator: Difference between revisions

Revision as of 15:00, 23 December 2009

Revision as of 15:10, 23 December 2009

Line 158:

== Scharr operator ==

The Sobel operator, while reducing artifacts associated with a pure central differences operator, does not have perfect rotational symmetry. Scharr looked into improving this property and found that using the following kernels could produce better results <ref>Kroon, 2009, Short Paper University Twente, [http://www.k-zone.nl/Kroon_DerivativePaper.pdf ''Numerical Optimization of Kernel Based Image Derivatives'' ].</ref> <ref>Scharr, Hanno, 2000, Dissertation (in German), [http://nbn-resolving.de/urn/resolver.pl?urn=urn:nbn:de:bsz:16-opus-9622 ''Optimal Operators in Digital Image Processing'' ].</ref> <ref>B. Jähne, H. Scharr, and S. Körkel. Principles of filter design. In Handbook of Computer Vision and Applications. Academic Press, 1999.</ref>:

<math>

Line 172:

\end{bmatrix}

</math>

There are more rotational invariant kernel solutions for other image scales, but the highest angle accuracy can only be obtained by using larger 5 x 5 schemes, <ref>Kroon, 2009, Short Paper University Twente, [http://www.k-zone.nl/Kroon_DerivativePaper.pdf ''Numerical Optimization of Kernel Based Image Derivatives'' ].</ref>

==See also==