27.07.2013 Views

AIR Tools - A MATLAB Package for Algebraic Iterative ...

AIR Tools - A MATLAB Package for Algebraic Iterative ...

AIR Tools - A MATLAB Package for Algebraic Iterative ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Chapter 2<br />

Theory of Inverse Problems<br />

and Regularization<br />

Inverse problems arise in many applications in science and technology. Examples<br />

where inverse problems are found could be in medical imaging, where it is used<br />

e.g. in CT scanning, in geophysical prospecting or image deblurring. We will in<br />

this chapter introduce some of the fundamental concepts of inverse problems.<br />

We will first introduce the concept of an inverse problem and describe what<br />

defines an ill-posed problem. Then the important tools of SVD and the discrete<br />

Picard condition is defined followed by a few examples of spectral filtering.<br />

Finally we will give a short description of semi-convergence <strong>for</strong> iterative methods<br />

and define the concept of resolution limit.<br />

2.1 Discrete Ill-Posed Problems<br />

Inverse problems arise when we need to compute in<strong>for</strong>mation that is either<br />

internal or hidden. In the <strong>for</strong>ward problem we have a known input and a known<br />

system and we can then compute the output. In the inverse problem the output<br />

is often known with errors and we then have to compute either the system or<br />

the input, where the other one is known. For the linear problems we let the<br />

system be represented by the matrix A ∈ R m×n , the output as the right-hand

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!