Improvements to channel cleaning

CASA Modification Request 3C108, November 2007

1. Introduction

The clean algorithm searches for the highest residual point in an image, and then subtracts a beam pattern appropriately scaled, center at the location of the highest residual point. At the present time, clean and mosaic search over the entire image cube for the highest residual point. If one of the channels is noisier than the other channels, for example, the clean algorithm will usually find the highest residual point in this channel and perhaps never get to the other channels. It is more robust to channel differences if the clean algorithm searched on one plane at a time for the highest residual point.

The above comments apply to the mosaic and wproj tasks as well.

The above comments probably apply to the cleaning of Stokes parameters. Each Stokes parameter (perhaps with Q and U excepted) should be cleaned one by one, instead of globally in the Stokes cube of (I,Q,U,V).

2. Background

3. Requirements

The implementation is probably a simple change in the search algorithm looping in order to find the next residual point to subtract. It is unclear if this reording will significantly change the speed performance.

If there is a significant loss of efficiency when going to a channel by channel cleaning, then there should be an option to clean globally or locally (one channel at a time). The global clean is useful for non-interactive scripts, the local clean is needed for interactive clean. However, if the channels have significantly different statistical properties, local cleaning is better.

[BG] It is only essential to change the order for interactive cleaning. Non-interactive cleaning can be left for a later development if difficult.

4. Design

5. Deployment Checklist

6. Test Plan

6.1 Internal Testing

6.2 Sponsor Testing

Can be tested easily since the reordering of the cleaning will be obvious

6.3 Integration/Regression Tests

Probably included in regression tests already

6.4 Testing for Scientific Validity

Some deep cleans of spectral line data should be compared using the clean before and after the suggest implementation. There may be some differences in the images because of the non-linear properties of clean (depends on the order that clean components are removed), but they should be within the noise level.


APPROVED: I acknowledge that my request is fully contained in this MR, and if the CASA development group delivers exactly what I specified, I will be happy.

ACCEPTED: I acknowledge that I have validated the completed code according to the acceptance tests, and I am happy with the results.

Written - - - - -
Checked - - - - -
Approved by Scientific Sponsor - - - - -
Accepted/Delivered by Sponsor - - - - -

  • Use %X% if MR is not complete (will display ALERT!)
  • Use %Y% if MR iscomplete (will display DONE)

Discussion Area

-- NicoleRadziwill - 30 Oct 2007

-- EdFomalont - 01 Nov 2007

-- EdFomalont - 05 Nov 2007
Topic revision: r3 - 2007-12-07, BrianGlendenning
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback