IRIS-Beginner Tutorial

8/13/2019 IRIS-Beginner Tutorial

1/17

IRIS-Beginner Tutorial

This tutorial provides a step-by-step procedure for processing digital images obtained in raw format toyield instrumental magnitudes which are used in the next step in our tutorial: processing the data usingExcel. This process assumes that, in addition to a series of sequential star field images, dark frames,flat frames have been taken. This tutorial also assumes that you have already copied the images from

your camera to some directory on your computer this greatly speeds up processing!.

"#"$ is a powerful image processing utility with a command-line editor and a graphical interface %&"!.'e attempt to highlight both methods below.

"n this tutorial you will

1. Initialize IRIS

2. Check raw images

3. Load and conert images

!. "er#orm Bias and $ark su%traction then &lat diision

'. (lign and stack images

). Select the green channel o# the images stack #ile

*. "er#orm (+erture +hotometr,.

. +tion/ (nal,ze each image instead stacking them

(fter these steps, the instrumental magnitudes are entered into a spreadsheet which will yieldcalibrated magnitudes which can then be submitted to (()$*.

The spreadsheet performs both:

- the color correction to transform the green +$# response to the ohnsons ) standard

- the compensation of the differential atmospheric extinction between the stars of the image

This tutorial was written by %/ara, bikeman, and #oger 0ieri with editorial changes by bkloppenborgand screenshots by bikeman.

0ote1 ersion/'e recommend using the latest version of "#"$ 1.12 as of the time this documentwas reviewed: (ug-3455! because it implements several new features that make it easier to work withraw images.

0ote2 la+sed Time/'hen processing a series of images, "#"$ will pause for some time to read,transform, calculate and save the images. The time needed for image registration could be long:several 54 seconds to minutes per image depending the image si6e and the computer. *nly the finalresult of these calculations will be displayed as the current image or values in *utput 7ox dependingthe operation.

0ote3 Linu4/ 8or inux users: This tutorial was tested under &buntu 54.49 using 'ine 5..3; with

"#"$ 1.12. The only thing that does not work is the drag-and-drop of images files from 'indowsExplorer. )E#T#(' command in order to load a series of raw images files.

0ote! 0eed #or $ark/+ifferential dark currents that generate a 8ixed 0attern >oise 80>! aredepending of the technology level of the sensors, they are not determined by the physicsfundamentals but #andom >oises: read noise, shot noise, are ?! *n recent sensors, like +$#=/*$ 59 bits, the dark noise is negligible in exposures shorter than 4 seconds. "n that case the darkprocess would not apply but "#"$ doesnt provide such option. *ne solution is to use @dummy@ imagesto see 9.5.3! for both bias and dark a simple constant equal to the system offset of the (nalog to+igital =onverter (+=! of the camera for the bias and @6ero@ for the dark!


2/17

0ote'/ IRIS Commands and 56I

"#"$ has been developed by =hristian 7uil during many years, at the beginning it was mostly acommand-line type of software, a kind of toolbox. (long years many tools have been added as newcommands and some connected to a %&" but not all. Existing commandsAtools are apparently nevermodified but new ones are Bust added. That means older applications are not impacted by newversions: when something doesnt work this is normally not a version issue.

"#"$ is down-loadable at: "#"$ exeThe definition of the newest commands is there. )ariousapplications examples are at end of the page.

The definition of the commands up to )1.1C can be find CA3455! at: "#"$ =ommands

(s tools are not all available through the %&" %raphical &ser "nterface! we will have to use bothways: the old command-line prompt and the %&".

The %&" functions are accessed through the /enu 7ar, the Tools 7ar below the /enu, and somefunctions are accessible via a =ontextual /enu at right mouse button click.

The /enu 7ar is explicit using texts, the Tools 7ar has buttons not so explicit ?

"n the following tutorial the call to the tools is shown in bold indented expressions like:7enu Bar/ &ile8Settings...

Tool Bar/ %utton with the camera +icture

*r, for the commands:

9 C0RTR(: input_file_name output_file_name number

The command shall be typed after the prompt @ D @ in the =ommand 7ox when "#"$ displays thatprompt if not it means "#"$ is computing and not available yet! The command name is shown in boldcapital letters but "#"$ is not case sensitive! then the small italici6ed letter words are input andAoroutput file names you should provide at your choice, none are imposed! and values of parametersere number is the number of images to process!.

0ote) Raw $SLR Image and IRIS/

(s a +$# doesnt provide info on sensor temperature or controls it the dark imagesshall be takenduring the imaging session with the same settings than the sky images. /ultiple dark images areneeded to reduce their read noise by averaging and then avoid to pollute the sky images. *ne solutionis to get 1 dark images before the imaging session and 1 at end, all at the operating temperature. 7utdepending the length of exposure and the sensor technology level, the dark process is not alwaysneeded To see note9 and $tep 9.3 !
http://astrosurf.com/buil/us/iris/iris.htmhttp://astrosurf.com/buil/us/iris/iris.htmhttp://astrosurf.com/buil/us/iris/reference/commands_iris_v5.58.pdfhttp://astrosurf.com/buil/us/iris/reference/commands_iris_v5.58.pdfhttp://astrosurf.com/buil/us/iris/iris.htm


3/17

"#"$ needs a master-%ias, the code being not flexible on that point To see 9.5! "f a formal bias isused it should be made from short dark exposures to avoid dark currents to develop. ( masterFbiascould be reused for many sessions.

&latsare mandatory to compensate the vignetting of lenses To see $tep ., 9. and 9.1! ike for adark multiple flat images are needed to reduce the read and shot noise by averaging. 7ut "#"$ doesntuse flat-dark but apply only a bias to it. That means the flat images are expected taken at high enoughshutter speed at which dark current cant develop. +$# and photo lenses having large 8*), theirflats are difficult to obtain. "ts very difficult to ensure the illumination A reflectance of the target isuniform below a couple of percents 5G is 4.45 mag ??? ! ( master-flat could be reused for manysessions.

Sk, Images: ( long enough sum of exposures is needed for averaging the scintillation of the stars. (typical series sum is about ;4 seconds but it should be noted the scintillation level is very variabledepending the sky condition. The variability due to the scintillation is well quantified by calculating the$tandard +eviation $+! of the star magnitudes of a series of images or from a series of stacks.

( series of images could also be affected by @8lying $hadows@, any faint small cloud passing by,bubble of wet air... all not visible in a dark sky. To detect it a classical technique is to make severalseries and check the coherence of the averages of the series and relevant $+.

"n case of operation without tracking eg. on tripod! each exposure would be limited to 1H51 secondsdepending the focal length being used, the "$*, the stop, the defocus and the star magnitude. ( goodsolution is to get 1 series of 1 images then stack each series into one image using "#"$. >ext makethe photometry of each of the five resulting images, then compare the results. "f one image stackstrongly deviates from the others you could check the images of that series into details To see $tep 3!Then eventually eliminate that series or images! if something wrong is identified. (t end, after thespreadsheet processing, calculate the average and $+ of the 1 series, report that average, the $+ orthe $tandard Error! to =iti6en $ky.

Registration;Stacking : The registration of images with "#"$ could take a lot of time for images havingmany pixels like such of recent +$#. This is also very depending of the registration mode beingused. "f the lens is a 6oom there is a risk of serious geometry distortion of the image. "n this case the@%lobal matching@ and @Iuadratic@ options shall be used but the registration takes a lot of time. "f a

single long focal lens is used the distortion is probably very low. Then each series should be shotquickly, doing it within one minute should keep the star field rotation low enough. "n this case the @*nestar@ mode should be usable and much faster. "f there is a significant mismatch of the stars at theperiphery either @Three 6ones matching@ or @%lobal matching@ with @(ffine@ should be applied. @=ubic@is to be used in case of strong distortion.

(nother way than stacking, when few stars are involved, is to make the photometry of each image andthen average the results. 8urther its possible to make a dee+er statistic anal,sis. This process isdescribed under the $tep C option.

IRIS #ile series num%ering/

"n all cases the numbering of file sequences is generated A imposed by "#"$. The way this numberingis done is not compatible with most camera format and anyhow it starts at 1 #or each newse


4/17

Links/

=hristian 7uils "#"$8rom where to download "#"$ and find a lot of information

"#"$ =ommands The documentation of "#"$ commands up to version 1.1C pdf!

+$# 'ikipedia, general presentation with many links and table of sensor si6es

$ensor Types /ore on sensors, $i6e, $># vs si6e...

=8( A 7ayer =olor filters used in +$#

"mage >oise 80>, read, shot, quanti6ation... with links to more details

%amma The standard non-linear coding of image luminance 0E%, s#%7, >T$=...!

+ave =offins $ite +=#(', the raw image file converter used by many imaging software

dcraw.exe =hoice of dcraw.exe compilations for various systems

=$ky $preadsheet To finish the processing, color transform and extinction compensation

=alibration $tandardist of =omparison stars and related data
http://astrosurf.com/buil/us/iris/iris.htmhttp://astrosurf.com/buil/us/iris/reference/commands_iris_v5.58.pdfhttp://en.wikipedia.org/wiki/DSLRhttp://en.wikipedia.org/wiki/Image_sensor_format#Sensor_sizehttp://en.wikipedia.org/wiki/Color_filter_arrayhttp://en.wikipedia.org/wiki/Image_noisehttp://en.wikipedia.org/wiki/Gamma_correctionhttp://www.cybercom.net/~dcoffin/dcraw/http://www.rawness.es/?lang=en/http://www.citizensky.org/content/calibration-intermediatehttp://www.citizensky.org/content/calibration-standardshttp://astrosurf.com/buil/us/iris/iris.htmhttp://astrosurf.com/buil/us/iris/reference/commands_iris_v5.58.pdfhttp://en.wikipedia.org/wiki/DSLRhttp://en.wikipedia.org/wiki/Image_sensor_format#Sensor_sizehttp://en.wikipedia.org/wiki/Color_filter_arrayhttp://en.wikipedia.org/wiki/Image_noisehttp://en.wikipedia.org/wiki/Gamma_correctionhttp://www.cybercom.net/~dcoffin/dcraw/http://www.rawness.es/?lang=en/http://www.citizensky.org/content/calibration-intermediatehttp://www.citizensky.org/content/calibration-standards


5/17

Step 1 - Initialization:

1.1 Files Settings:

7enu Bar/ &ile 8 Settings...

=hoose: - your working path where you have =*0"E$ of your raw images files!,

- the script path if you want to use some scripts.. ! and

- the file type for images produced by "#"$ eg. 0"= !.

1.2 Camera Settings:

Tool Bar/ %utton with the camera +icture

*pen =amera $ettings and leave the defaultvalues except:

- 7inning 5x5,

- =amera model select appropriate!

- #aw interpolation method @inear@,

- 'hite balance >*T selected.

>ote: @=amera model@ is used to determine the color order of the =olor 8ilter (rray and the

suffix extension! of the raw image files then other types are not recogni6ed by "#"$! /ost=anon are #%%7, >ikon 7%%#. This information is available as output of the +=#(' rawimage file converter. "f your camera is not in the list choose first a similar one from the samebrand. "f the problem remains visit the site of +ave =offins dcraw.exe the converter used by"#"$! &sing @dcraw -i -v "/%Fxxxx.=#3@ =anon! as a 'indows command shows the imageinfo including the =8( order. >ear 144 camera types are covered.


6/17

Step 2 - Checking Images:

"#"$ has some simple graphical tools that enable us to check images quickly before processing them.This is useful for optimi6ing the shooting parameters, framing, avoiding saturation... "t could be usedfor further analy6ing the images when problems occur.

The raw images can be visuali6ed simply as a =8( A 7ayer image using the raw loading:

7enu Bar/ &ile 8 Load a R(: #ile...

Slider Bo4/ (uto %utton

$lider 7ox: #ange option!

#aw images are coded linearly on 53 or 59 bits and are not in a shape to be properly visuali6ed.&sual images like Bpg! are coded, shaped, with a low power function not far a log.. ! following thestandard 7T J42 to be properly displayed %amma!.'ithout such standard gamma raw images lookvery dark on usual displays, only the brightest stars are visible if not saturated. Even applying agamma would not be enough as our images go very deep, fainter stars recorded are ; magnitudesbelow the brightest non-saturated, this is a ratio of 314, too much for any display. "#"$ provides a toolthat enables us to properly clip the image at an adBustable level and also control the black level. This isthe @Threshold@ $lider 7ox that pop up at start of "#"$ if not use: Tool Bar/ Button with two arrows!

- The upper slider sets the clipping level of the image,

- The bottom slider adBusts the black level.

The @(uto@ button starts the automatic search of thebackground level of the image and applies clipping andblack level shift defined through @#ange@

The @Threshold #ange@ values are such of the +$#:

- 9421 A 4 for a 53 bits

- 5;C A 4 for a 59 bits.

@+elta i@ determines the clipping level above the background plus offset.

@+elta ow@ elevates the display black level by its value, this better shows thefaint stars and the noise.

8or our 53 bits samples +elta i K 144 and +elta ow K 544 is fine. Thenduring your image checking you could change the clipping and black levelsusing the sliders at will.

(ll this onl, a##ects the dis+la, not the image data.

There are other #unctions under/

7enu Bar8iew8...

that could %e used to either generate a standard gamma a logarithm or an, otherad=ustments %ut all are a++lied to the data and shall neer %e used #or +hotometr,.
http://en.wikipedia.org/wiki/Gamma_correctionhttp://en.wikipedia.org/wiki/Gamma_correctionhttp://en.wikipedia.org/wiki/Gamma_correction


7/17

2.1 Checking Saturation, ange, !ackgroun", #oise...

(nother simple tool is available through the mouse. "f you want to check a star Bust draw a boxaround it with the left button down, then click right, a dialog box pops up, click @$tatistics@

(nother box shows you various parameters of interest.

7ouse >9 draw %o4 >9 click right >9 dialog %o4 >9 Statistics

Saturation check/The @/ax@ value is such of the highest pixel. 'ith a ==+ 53 bits with anti-blooming we should probably not go higher than 3144H444. "n a 59 bits =/*$ the sensorwould saturate about 59444 (+&s at "$* 544 and then, at higher "$*, the (+= range shouldclip the image at 5;C (+&s but much less electrons, limiting the dynamics... !

Background Leel/The @/edian@ value shows it. To get an exact value the box should bedrawn in a dark area. Then the @/ean@ is a more accurate value. 'ith =anon +$# that level

includes the system offset of the (+=, in general 53C for 53 bits and 5439 for 59 bits.That value shall be subtracted from the image data to determine the true background level.

0oise leel/ 'hen used on a dark the @sigma@ shows the noise level %aussian read noiseand possibly some dark current impulse noise! "f used on images it could includes some othersignals.

2.2 Checking $efocus, %rail

7ouse >9 draw %o4 >9 click right >9 dialog %o4 >9 ?

..Cro+/a box drawn around a star and then using the 6oom buttontwo buttons before @L5@ ! permits to visuali6e the star pixels intodetails, to check the defocus, the trail.

This is a @7ayer@ image, the #%%7 sensor arrangement, but in 7M'.

"f you point a pixel with the mouse its value is shown at the bottom-right of the window after its coordinates.


8/17

2.& Setting the 'hotometr( )perture

7ouse >9 draw %o4 >9 click right >9 dialog %o4 >9 5rowth cure

This function shows the photometry error infunction of the si6e of the inner circle of thephotometry tool. "n our test set the photon count iscompleted below 4.45 mag for a 53 pixels radius.

( ; pixel radius would result in losing photonscorresponding to an error of 4.3 mag or 5JG ofthe flux. To be applied in $tep J.

2.* Checking the "ark images

This can be done with the @istogram@function applied to a raw dark image:

7enu Bar/ Load a Raw #ile...

7enu Bar/ iew 8 @istogram

"lot 7enu Bar/ +tion 8 (4is Setu+


9/17

Step & - +oa"ing an" Conerting a Image Files:

&.1 pening the loa"ing !o/

7enu Bar/ $igital +hoto 8 $ecode R(: #iles...

This will open up a box in which files should be dragged-and-dropped from 'indows Explorer.

$rag and dro+ ,our raw image #ileseg. for =anon: "/%F>>>>.=#3 files! here.

&.2 Conert ra image files to CF) files

Bo4 Button/ A C&(A

This dialogue lets you specify an output generic name.

(fter dropping in your images, press the @--D=8(@ button to decode the images inux userssee the note here below!

&.& )ppl( this processfor:

- star field images e.g. use the name @img@!,

- dark frames suggested name @dark@!,

- flat field images suggested name @flat@! and

- bias frames @bias@ optional!.

>ow you should have a series of images, darks, flats and optionally bias frames in the workingdirectory at the format you configured in 5.5

>ote: "f you are running "#"$ under inux also works under 'indows!, open up the commandline input dialog box, the button to the left of the camera button, and type at the prompt @D@

9 C0RTR(: input_file_name output_file_name number

'here @input_file_name@ is the image file generic name from your camera and@output_file_name@ is the generic names of files that "#"$ creates with the =8( order definedby 5.3 and file type defined in 5.5 7oth names do not include the sequence numberingimposed by "#"$! nor the file suffix.

The full inuxA'indows name of input files would be: @"/%F422>.=#3@ for =anon for example,where @>@ is a series of numbers that starts at @5@ and ends at @number@ without leading 6eros.This is in general not compatible with the camera numbering format and you would have toconvert the numbering of your files.


10/17

Step * - 'reprocessing 0 !ias, $ark, Flat:

&nlike ==+ cameras, +$# cameras dont provide the user with the ability to control the sensortemperature and apply a previously computed master dark and bias combination to the images.7ecause of this, one should take dark shots during the imaging session which will include the biaspattern of the camera. (lthough the +$# dark frames will account for the bias, "#"$ requiresseparate bias offset! frames. 7ecause of this, we can either create a formal bias frame or a fake@dummy@! bias frame, both work equally well, but one of the two must be completed.

(s said in >ote9 recent sensors have very low differential dark currents and bias pattern, forexposures shorter than 4 sec dark and in any case bias! can be replaced by dummies.

*.1 !iasffset Frame

*.1.1 Creating a Formal !iasTo generate a 7ias /aster 8rame go to:

7enu Bar/ $igital "hoto 8 7ake an o##set...

- Type the appropriate generic name the one introduced in $tep , e.g. bias!

- The number of bias images,

Bo4 Button/

- Then select the @command line@ button the one on the left from @camera@ button! andtype at the prompt:9 S( master-biasor any name you would like! or 7enu Bar/ &ile 8 Sae..

The master-biaskee+s alidfor a given +$# as long as its electronics remains in goodcondition, a check every couple of months would be enough. "t is "$* depending.

*.1.2 Creating a "umm( !ias

$ome +$#A+$= cameras use a @system offset.@ This offset is built into the cameras analogto digital convert (+=!. This offset will take the black level ideally 4! and encode it as a highervalue usually 53C for 53-bit cameras, 31; or 5439 for 59-bit cameras!. the result is thatgenerated values can be less than the offset 6ero position or, in other words, negativebrightnesses are permitted? eg. %aussian read noise! "#"$ doesnt handle this parameter,even if the camera types are identified, therefore we need to determine it for generating theartificial bias.

The easiest way to do this is to load one of the dark frames and look at its black level. Tomeasure the black level use the command prompt box and type:

9 ST(T

The output window will show an average value of the whole image. &se this value as the black6ero-point reference.

"f you would rather not use the command prompt, you can select a blackened portion of theimage using the mouse. To do this left click on the area, draw a box, right click and select@statistics.@ "#"$ will show the average value in the boxed area. &se this value as the black6ero-point.


11/17

>ow we create the artificial bias frame. &se the command prompt box and type:

9 &ILL alue

'here @value@ is the black 6ero-point from above. >ow, save the master bias by typing:

9 S( master-bias$uch dummy master-bias is valid for ever.

ow we need to create a master dark:

7enu Bar/ $igital "hoto 8 7ake a dark...

Type in the box the generic name, the offset imagethe master bias frame you Bust created!, and thenumber of images in the dark sequence. $electeither the mean or median method they work equallywell, median has the property to eliminate anyinfluence from extreme values!, and press *N. Thentype the command:

9 S( master-"ark

or choose:

7enu Bar/ &ile 8 Sae...

To create a Adumm, darkAif your +$# is recent and your exposure short enough, Bust usethe same process as in 9.5.3 but apply 9 &ILL D

*.& 3aster Flat:

The flat process is mandatory for eliminating the vignetting of the lens. )ignetting as high as4H94G are common at large aperture and could induce large differential magnitude errorsdepending location of the reference star and the target star.

#epeat the same process as used for darks:

7enu Bar/ $igital "hoto 8 7ake a #lat-#ield...

Except before saving type at the prompt:

9 5REF&L(T in the command box to normali6e the red and blue =8(-flat.

Then save the file as @master-flat.@


12/17

That master-#lat kee+s alideach time you use the same +$# and ens at the sameaperture, same focus, same focal length for a 6oom. =heck that no large dust has polluted thelow-pass filter that is few millimeters in front of the sensor, this is the only level a dust could bean issue if you use a large lens aperture 8A9 or less! (ny flat would not correct it properly.

*.* 4ot 'i/els $etection an" ecor"ing:

This function automatically detects hot pixels above a threshold level. $uch defective pixels

couldnt be processed properly through dark or flat. Their coordinates are recorded in a file forfurther processing of the images replacement by surrounding average! "t might take someexperimentation to determine the threshold value you need, but a good starting value at"$* 544 is about 344 (+&s for a 53-bit ==+ cameras and 144 for a 59-bit =/*$ camerasoffset removed!. That threshold should be more or less proportional to the "$* being used.The number of hot pixels could be 14 in the first case or 54 in the second very depending ofthe sensor quality class.

oad the processed master dark frame created in $tep 9.3 , offset removed by "#"$! and typethe following at the command prompt:

9 &I0$F@T cosme number

'here @number@ is the threshold value you have selected and @cosme@ is the name of the filewhere the results are recorded you could chose any you like!. =heck the *utput 7ox it shouldopen automatically!. "f there is a number of hot pixels not detected, change the thresholdaccordingly.

"ts possible to differentiate the random noise, the impulse noise and the hot pixels using thehistogram function, the threshold should be well above the %aussian >oise distribution, thenthe stack of +ark "mpulses and below the isolated peaks on the right corresponding to the hotpixels.


13/17

Step 5 - )lignment of Images:

5.1 CF) conersion to 6!:

7enu Bar/ $igital +hoto 8

Seote;,

you shall load first the first image and draw a boxaround the star!

5.& Stacking the images:

7enu Bar/ "rocessing 8 (dd a se


14/17

Step 9 - Selecting the green channel:

The recent versions of "#"$ i.e. 1.1J or newer! include a quick and easy way to pull out the #%7channels:

Load the stacked image #ile eg. imgFstk ! if notalready in memory and go to:

7enu Bar/ $igital "hoto 8 R5B Se+aration...

Then Bust type in the dialog box:

- the names of the color channel files:

e.g. final-r, final-g, final-b!

- =lick *N


15/17

;.1 'hotometr( Setup

7enu Bar/ (nal,sis 8 (+erture +hotometr,...

'ith our sample data, you should probably use thesi6e of the reading apertures shown at left, but for yourown data you may need to change these si6es so youmay wish to review how to si6e a photometry aperture

To see 3.!The inner circle is the area in which the star photon-count is done, the external annulus is used to measurethe average or median level of the background that isnext subtracted from the inner circle pixel values.

The intermediate annulus is used as a guard to avoidthe halo of the star pollutes the background. "t could beused to eliminate a nearby companion star. The inner

circle should well intercept most signal from the star and by the way be significantly larger thanthe largest visible star footprint. The same settings shall be used for all stars of a given image.

The background area should be large to minimi6e the involved noise if no background star fall

in it.

The @/edian 7ackground@ method is less sensitive to faint background stars pollution but isless accurate integer value! than the mean floating point value! ( small rounding error due tothe integer mode will be multiplied by the number of pixel in the inner circle. *ne half (+& errorwould result in 331 counts deviation of the star intensity in our example ?


16/17

%o to:

7enu Bar/ (nal,sis 8 7agnitude Constant...

Set the Constant

>ow all readings will be made relative to that

=omparison $tar but neither color nor extinctioncorrected, there are only approximate andprovisional.


17/17

Step 7: 'hotometr( of all images ithout stacking 8ption

"f you prefer to make the photometry of all your images and not stack them, Bust after thepoint 1.5 apply the following sequence #%7 separation:

7enu Bar/ $igital "hoto 8 Se

IRIS-Beginner Tutorial

Documents

Transcript of IRIS-Beginner Tutorial