Summary: Rice University ELEC 301 project looking at the use of Weiner filters in the deconvolution of multiple noisy astronomical images into a single, clean image.
In the realm of image processing, one of the problems that is frequently encountered is the problem of deconvolution; given a noisy, blurred signal, how do we estimate and remove the noise and distortion and thereby obtain a clean copy of the original signal? The processes used given a single input to the deconvolution (single-input single-output deconvolution or SISO-D) are well studied, and a variety of techniques have been developed to cope with this problem. However, the realm of multi-input single-output deconvolution (MISO-D) is still being explored, and new strategies are being developed for optimal deconvolution given multiple inputs. Our goal is to use a basic SISO-D technique, the Weiner filter, applied to multi-input data to obtain a clean copy of the orignal signal for a set of astronomical images of the globular cluster Messier object 3, and to show the overall strategy used in current MISO-D image processing techniques.
| Messier Object 3 |
|---|
![]() |