# Cleanlens


*Under this identifier are defined some parameters to retrieve the shape of the source knowing a pixel-frame of the image.*

Here the program reads CCD pixel-frame and using the equation of the lens compute a pixel-frame of the source. For each point of the Image plane the program can compute the corresponding point in the Source Plane. Then for each pixel with multiplicity $n_{ij}$  of the source plane we then can attribute an intensity computed as the mean of the intensity of corresponding image pixels:

$\mathcal{I}^S_{ij} = \frac{1}{n_{ij}} \sum_{k=1}^{k=n_{ij}} \mathcal{I}^I_{ijk}$

The error of this reconstruction is given at position *ij* in the source plane by:

$e_{ij} = (n_{ij} - 1) \sum_{k=1}^{k=n_{ij}} (\mathcal{I}^I_{ijk} - \mathcal{I}^S_{ij})^2$


(the error estimate exists only if we have multiple images).

When using the [inverse]() mode, the program will minimize the estimate of the error ($e = \sum e_{ij}$). This *ring circle* method, nevertheless numerically different, was first described by [Kochaneck et al. 1989](http://adsabs.harvard.edu/abs/1989MNRAS.238...43K).




## cleanset int float


If `int = 0`, nothing is performed.

If `int = 1` then the program will compute from the image pixel-frame [imframe]() the corresponding source frame [sframe]() at redshift *float*. 

If `int = 2` then the program will compute from the image pixel-frame [imframe]() the corresponding source frame [sframe]() at redshift `float`, and will fit the parameters of a [source shape model]() to this image in the source plane.

If the [inverse]() mode is selected, an optimization of the _ring cycle_ form will be done. In this case the optimization will work only if multiples images or giant arcs are present in the image pixel-frame. It is strongly advise the user to have an a priori good guess on the mass model parameters.

Moreover it will create the pixel-frame: 
- `erreur.fits` : this frame is relevant only in the case of giant or multiple arcs, it computes the reconstruction error of the source pixel-frame for each pixels. 

- `imult.fits` : this pixel-frame shows the multiplicity of each pixel of the reconstructed source pixel-frame.


## c\_image filename

If `int = 1` for the cleanset keyword, filename is an ASCII file that contains a list of points in the format ($i$, $x_i$, $y_i$) in the image plane given in arcsecs relative to the [image reference points]() that defines the center of the observed image. The barycenter of those points is sent to the source plane and defines the source center. This value is used to compute the WCS keywords of the resulting source FITS file.

## imframe int filename


`int` specifies the format of the input image ( `int = 2` for ipx format, `int = 3` for FITS file).

`filename` is the name of the CCD frame where (multiple) gravitational images are present.


## psfframe int filename


`int` specifies the format of the input Point Spread Function (PSF)( `int = 2` for ipx format, `int = 3` for FITS file). 

`filename` is the name of the PSF frame, namely a Star Profile. The Star must be at the center of the frame. And the total intensity of the frame (sum of all the pixels) must be equal to 1 (PSF normalized). Note: the psf frame is used when using a deconvolve-inversion of the lensed images.


## sframe filename

`filename` is the name of the output source frame. This frame will correspond to the inversion of the [imframe]() frame (from Image Plane to Source Plane). It is written with the FITS format.

## wframe int filename


`int` specifies the format of the input image ( `int = 2` for ipx format, `int = 3` for FITS file).

`filename` is the name of the weight frame. This image is required to perform the optimisation.


## ncont int filename


`int` is the number of contours. 

`filename` is the name of the output frame: that is the CCD frame where only the pixels inside the contours (closed lines) have been kept. The idea is to limit the area of the frame, and keep only the interesting pixels.


## contour int filename


Defines contours in the image plane of one or several images of the source you want the shape in the source plane. 

`int` is the index of the contour for one image. First contour must be index by 1, second by 2, etc. 

`filename` is the name of the contour ASCII file for one image that contains a list of points ($i$, $x_i$,$y_i$) in the image plane given in arcsec relative to the  [image reference points]().


## subsamp int


It is possible to subsample the CCD frame to calculate with greater accuracy the source frame in the source plane. 

`int` is the subsampling parameter. Default is 1. If `int = 2` it will cut each pixel in 4 smaller pixels. Default value : 2.


## s\_subsamp int

It is possible to subsample the frame in the source plane as you can do in the image plane with the [subsamp]() keyword. If `int` is too high, you can get an image with undefined pixel values. Default value : 1.

## s\_n int

Defines the width and height of the resulting source image in pixels. Default value : 50 pixels.

## s\_xmin float

Defines the left (East) limit of the source plane image in relative arc seconds. Default value : -20 arcsec. This value is overwritten if a polygon in the image plane is provided with the [c_image]() keyword in the [cleanlens]() section.

## s\_xmax float

Defines the right (West) limit of the source plane image in relative arc seconds. Default value : 20 arcsec. This value is overwritten if a polygon in the image plane is provided with the [c_image]() keyword in the [cleanlens]() section.

## s\_ymin float

Defines the bottom (South) limit of the source plane image in relative arc seconds. Default value : -20 arcsec. This value is overwritten if a polygon in the image plane is provided with the [c_image]() keyword in the [cleanlens]() section.

## s\_ymax float


Defines the upper (North) limit of the source plane image in relative arc seconds. Default value : 20 arcsec. This value is overwritten if a polygon in the image plane is provided with the [c_image]() keyword in the [cleanlens]() section.


The following parameters should be set only if the `int` argument of the [imframe]() keyword is 0 (ascii format) or 1 (ipx format without scaling). It is strongly advised to used the ipx format: 2 or FITS format: 3, to avoid to define these parameters.


## pixel float

Sets size of the pixel in $x$ and $y$ in arcseconds.

## column int

Column to be selected (in a multi-column ascii pixel-frame file).

## header int

Number of lines to skip before the real beginning of the data.

## pixelx float

Pixel size in $x$ in arcseconds (if the pixel is not a square).

## pixely float

Pixel size in $y$ in arcseconds (if the pixel is not a square).

## xmin float

$x$ position of the bottom-left pixel (expressed in pixel units).

## ymin float

$y$ position of the bottom-left pixel (expressed in pixel units).