Unimap Documentation Page

The core of Unimap is an original and efficient implementation of the GLS map maker, which is an established noise removal algorithm, used by several other mappers. e.g. Madmap.

The GLS mapper is preceeded by a sequence of preprocessing modules which remove the following impairments:

> Offsets.
> Jumps due to cosmic rays.
> Initial drift due to calibration.
> Spikes (glitches) due to cosmic rays.
> Baseline drift.
> Systematic pointing error.
> Pixel noise.

The GLS mapper is followed by the PGLS and WGLS steps, which remove the distortion introduced by the GLS mapper.

A paper describing the whole data reduction pipeline is [4]. The PGLS/WGLS algorithm is described in [2]. The drift removal method is described in [3,5]. The pointing error compensation is described in [6]. The pixel noise correction is described in [7].


Click here to download a power point presentation of the data reduction pipeline (5.3.0).

Click here to download the latest Unimap User's manual.


[2] L. Piazzo, D. Ikhenaode, P. Natoli, M. Pestalozzi , F. Piacentini and A. Traficante: "Artifact removal for GLS map makers by means of post-processing", IEEE Trans. on Image Processing, Vol. 21, Issue 8, pp. 3687-3696, 2012.

Compact citation: Piazzo, L., Ikhenaode, D., Natoli, P., et al. 2012, ITIP, 21, 3687.

[3] L. Piazzo, P. Panuzzo, M. Pestalozzi: "Drift removal by means of alternating least squares with application to Herschel data", Signal Processing, vol. 108, pp. 430-439, 2015.

[4] L. Piazzo, L. Calzoletti, F. Faustini, M. Pestalozzi, S. Pezzuto, D. Elia, A. di Giorgio and S. Molinari: 'Unimap: a Generalised Least Squares Map Maker for Herschel Data', MNRAS, 2015, 447, pp. 1471-1483.

[5] L. Piazzo, M. C. Raguso, J. G. Carpio and B. Altieri: "Least Squares Image Estimation in the Presence of Drift and Pixel Noise", European Signal Processing Conference (EUSIPCO), Budapest, Hungary, August 2016.

[6] L. Piazzo, M. C. Raguso, L. Calzoletti, R. Seu, B. Altieri: "Least Squares Time-Series Synchronization in Image Acquisition Systems", IEEE Trans. on Image Processing, Vol, 25, Issue 9, pp. 4458-4468, September 2016.

[7] L. Piazzo: 'Least Squares Image Estimation for Large Data in the Presence of Noise and Irregular Sampling', submitted to the IEEE Trans. on Image Processing, copy available by the author, 2017.

Change log

Release 5.3.0 is a milestone since it was tested at the Map-Making Workshop held at ESAC in january 2013. The changes in the following versions are reported here.

* Release 5.4.0
- WGLS revised to improve automatic parameter setting.
- WGLS parameters changed.
- Code optimisation.

* Release 5.4.1
- Minor bugs fixed.

* Release 5.4.2
- Write problem with very big data set solved.

* Release 5.4.3
- Jump detection improved.
No need to reprocess if the image is good.
Can improve quality for images where some jumps were visible.

* Release 5.4.4
- Output Fits format revised.

* Release 5.5.0
- Option to remove initial samples from the timelines: hard way to remove the initial onset (calibration block memory) when the exponential fit does not work.
- Corresponding parameter top_skip added in the parameter file.
- Handling of input flags improved.
This version has a different parameter file.
Can improve quality for images with hard onsets and saturations.

* Release 5.5.3 (Linux) or 5.5.4 (Mac)
No changes, but about twice faster...

* Release 5.5.5
- Bugs fixed
- Automatic selection of some parameters
- Default parameters tuning
Quality of automatically produced images has improved

* Release 6.0.0
- Code optimized for speed and memory: twice faster...
- Automatic selection of more parameters
This is a major update.
This version has a different parameter file.
Processing is still essentially the same as in 5.5.0 but quality of automatically produced images is improved

* Release 6.1.0
- New start image for GLS added
- GLS convergence parameters revised
Works for some maps where 6.0 failed

* Release 6.2.0 - 20 Aug. 2015
- GLS Delta tuned for speed
- Drift and morphology improved
This version has a different parameter file.
A bit faster and slightly better than 6.1

* Release 6.3.3 - 20 Jan. 2016
- Parameter handling and file modified
- Pointing error correction added (prototype)
- Pixel noise handled in GLS (prototype)
This version has a different parameter file.
This version has important changes wrt 6.2.0
Click here for more info

* Release 6.4.0 - 07 Apr. 2016
- Parallel versions for GLS and PGLS
- Noise map computation
- Parameter settings and some code revised
Click here for more info

* Release 6.4.1 - 11 Apr. 2016
No software changes.
Only parameter file added to distribution.

* Release 6.4.2 - 21 Apr. 2016
Two bugs fixed.

* Release 6.4.3 - 17 may 2016
Minor changes - Parameter tuning.

* Release 6.5.0 - 30 may 2016
- Morphology map modifed.
- GLS Preconditioning implemented.
- Pixel noise estimation updated.
- Parameter and processing optimised.
This version has a different parameter file.

* Release 6.5.1 - 4 June 2016
- Bugs corrected.
- Morphology map restored to 6.4.3.
- Default params revised.
This version performs well on most reductions with the default parameters

* Release 6.5.2 - 8 June 2016
- No changes, just code cleaning.
Wrt 6.4.3, this version produces better maps in auto mode. Running time is almost halved.

* Release 6.5.3 - 14 June 2016
- One bug less.

* Release 7.0.0 - 5 April 2017
- Several bugs less.

* Release 7.1.0 - 26 April 2017
- Calibration blocks flagging revised.