ASTAP, the Astrometric STAcking Program    ASTAP Forum        Documentation     Version history   Checklist for solving   
             astrometric (plate) solver, stacking of images, photometry and FITS viewer

• Questions, feedback: ASTAP Forum
• Older versions are available at Sourceforge
• Version history
  
• Source code (Object Pascal, Lazarus/FPC)
•  

 

Donations are welcome:

  Documentation, contents:

• Introduction
• Program installation
• Program operation, stacking astronomical images
• Stack menu 
• Images, darks, flats, flat darks tab
  
• How to exclude poor images
• Satellite tracks
• Alignment menu tab
• Internal_alignment
• Astrometric_alignment
• Manual alignment
• Ephemeris alignment
  
• Results tab
• Stack method tab
• Stacking grayscale of colour stacking
• Stacking raw one shot colour images (OSC)
• Stacking LRGB
• Image stiching
• Blink tab
• Photometry tab
• DSLR photometry
• Inspector tab
• Mount analyse tab
• Live stacking
• Monitoring
  
• Pixel math tab
• Background equalization tool
• Popup menu
• Export FITS data to spreadsheet for analyses

• Viewer:
• Reported angle
• File menu
• Opening, preview files
• FITS thumbnail viewer
• Tools menu:
  
• Add SIP coefficients to the header
• Batch conversion
• CCD inspector
• CCD inspector, tilt
• HFD_2d_contour
• HFD_diagram
• Median_background_values
• HFD values
• Unroundness of the imaged stars.
  
• Show distortion
  
• Aberration Inspector
  
• Photometry calibration
• SQM measurement
• Star annotation and photometry
• Asteroid annotation
• Deep sky annotation
• 
  
• View menu
• Popup menu
• FITS tables 
• Other popup menus
  


• Solver, usage as astrometric solver and command line options
• Conditions required for solving
• Test if an image is solvable (full blind solving)
• ASTAP native command line 
• Performance blind solving
• Popup notifier
• Usage as a solver with other programs or as PlateSolve2 substitute
• Using with APT,  Astro Photography Tool
• Using with BackyardNikon, BackYardEos
• Using with CCDCIEL
• Using with ModelCreator
• Using with NINA, Nighttime Imaging ‘N’ Astronomy
• Using with SGP, Sequence Generator Pro
• Using with SharpCap
• Using with Voyager
• Installation of the external astrometry.net solver 

• Appendix 1 stack process
• Appendix 2 Why using flat-darks or bias
• Appendix 3 the star database 1476 and 001 format.
• Background info, ASTAP star pattern recognition and astrometric (plate) solving algorithm.

ASTAP introduction

ASTAP is a free stacking and astrometric solver (plate solver) program for deep sky images. In works with astronomical images in the FITS format, but can import RAW DSLR images or XISF, PGM, PPM, TIF, PNG and JPG  images. It has a powerful FITS viewer and the native astrometric solver can be used  by CCDCiel, NINA, APT or SGP imaging programs to synchronise the mount based on an image taken.

Main features:

1. Native astrometric solver, command line compatible with PlateSolve2.
2. Stacking astronomical images including dark frame and flat field correction. 
• Filtering of deep sky images based on HFD value and average value.
• Alignment using an internal star match routine,  internal astrometric solver.
• Mosaic building covering large areas using the astrometric linear solution WCS or WCS+SIP polynomial. 
• Background equalizing.
3. FITS viewer with swipe functionality, deep sky and star annotation, photometry and CCD inspector.
• FITS thumbnail viewer.
  
• Results can be saved to 16 bit or float (-32) FITS files.
• Export to  JPEG, PNG, TIFF( ASTRO-TIFF), PFM, PPM, PGM  files.
• FITS header edit.
• FITS crop function.
• Automatic photometry calibration against Gaia database, Johnson -V or Gaia Bm
• CCD inspector
• Deepsky and Hyperleda annotation
• Solar object annotation using MPC ephemerides
• Read/writes FITS binary and reads ASCII tables.
  
4. Some pixel math functions and digital development process
5. Can display images and tables from a multi-extension FITS.
6. Blink routine.
7. Photometry routine
8. Available for MS-Windows 32 & 64 bit,  Linux 32, 64 bit, MacOS 64 bit, Raspberry-Pi Linux 32 and 64 bit.

Stacking of images:

Stacking of astronomical images is done to achieve a greater signal to noise ratio, prevent sensor saturation and correct the images for dark current and flat field. Additional imperfect images due to guiding, focus  problems or clouds can be removed.

This is a screen short of the stack menu. It contains several tabs for the file list and settings. File can be sorted on quality and values.The image can be visually inspected in the viewer by a double click on the file or using the pop-up menu.



Program requires FITS images or RAW files as input for stacking, but it can also view 16 bit PGM /PPM files, XISF files or  in 8 bit  PNG, TIFF or BMP files. For importing DSLR raw images  the program DCRAW from David Coffin or LibRaw is used.

• Stacking methods: average and sigma-clipping-average.. Internal calculation using floating point numbers.
• Simple and intuitive user interface.
• Automatic saving of selected options and files.
• Can create master files for dark and flat & flat-darks to reduce processing time.
• Limited memory use, independent of the number of images stacked.
• Bayer algorithm for DSLR/OSC cameras

For stacking the internal routine compares the image star positions to align.



Astrometric Solving:

ASTAP can be used as astrometric solver to synchronise the telescope mount position with center position of an image taken with the telescope. Existing images can be solved to annotate, for photometry or the measure positions of unknown objects.

The ASTAP solver aims at a robust star pattern recognition using the catalog star coordinates in Equinox J2000. The solution is not corrected for optical distortion, refraction, proper  motion  of  stars and other minor effects all to be very minor.

The process astrometric solving is often referred to as a "plate solve". That was a correct description in the past, but in modern times there are no photographic plates involved in the process.


Back to index

Program installation:

MS-Windows:


Linux installation:


MacOS installation:




Program operation, stacking astronomical images:

The purpose of the stacking routine is to combine astronomical images to reduced noise and to flatten the image.

 Ideally you should have collected

• 1a) Several light frames. Images of deep sky object unprocessed.
• 1b) Several dark frames of the same temperature and exposure as the light frames. A dark frame is a frame that represents an exposure done in total darkness. This signal includes the bias signal, but also includes any dark current charge accumulation, and thus any dark current noise that exists within the dark current signal. For darks suburban area (SQM=20.4) you should take about the same amount or more darks as lights. For a light polluted area you could take less darks then lights since the noise in the lights generated by the sky background will be abundant.
• 2a) Several flat frames. A flat frame is a frame that represents the field flatness and taken from an uniform light source. Ideally with a significant signal level. This to compensate for vignetting and dust particles. Vignetting can greatly darken the corners of your image and have to be compensated.
• 2b) Flat dark frames or bias frames ideally of the same temperature and exposure duration as the flats. Since flats are taken with very short exposure times, either flat dark of bias image of almost zero seconds will do. See also why flat-darks

Only light frames are essential.

The automatic stacking process in ASTAP goes through the following steps:

1. The flats will be combined to an average and the combined average flat-darks will be subtracted to  have a near ideal presentation of the vignetting called the master flat frame.
2. The darks will be combined to an average master dark.
3. From each light frame the master dark will be subtracted to extract the pure deep sky signal.
4. Each light frame will be flattened by dividing it by the master-flat resulting in the corrected light frames.
5. The corrected light frames are combined to the final image using the average or sigma clip mean (to remove outliers as satellite tracks) method. 

Steps 3,4,5 are done in memory. No intermediate results are stored on disk.

It is possible to mix different exposure times but it is not recommended.


Operation of the stacking program

Start the ASTAP program.

Call up the stack menu window using the  ∑   button.

a) Select frames
In tab images select the lights. In tab dark select the corresponding darks.  Select in tab flats the flat-field images called flats and in tab flats-darks the flat-darks/bias frames. In most cases you could select all frames in tab images. The program will move the fames to the corresponding tab during analyse. The light and dark should preferable have the same exposure time and temperature. The flats should have the same exposure time and temperature as the flat-darks.

b) Analyse and remove bad frames
In tab images (for the light frames), press analyse and remove manually any poor image. Poor images can be detected by a too high HFD (Half flux diamater stars), low number of stars or high background value( by clouds) . Loss of tracking could result in too low HFD value. If required inspect each image by double on the file name. The list can be sorted by clicking on the corresponding columns. Using the pop-up menu selected bad frames can be renamed to *.bak for deletion later.

c) Set parameters in tab stack method
In tab stack method select the stacking method, average or sigma-clip-average. For OSC camera images  select "Convert OSC images to colour". Select the correct Bayer pattern (4 options). Test the required pattern first in the viewer with a single image. The source images should be raw (gray)  without colour produced by  astronomical camera's.

d) Set parameters in tab alignment
Leave this to the default star alignment.

e) Classify by
Leave all check marks initially unchecked. (This is an option to select automatically a master dark with the correct temperature and exposure time for the lights. Same for  master flat selection  based on filter used both in the light and flat.)

f) Press the  Stack (..) button.
The darks and flats & flat-darks will be combined in a master dark and master flat frame. Then the program will combine the light frames to the final image and save it automatically to FITS. This will take some time.

g) Export
The stack result will be saved as FITS. The program keep a record of all results in tab Results. Stretch the image as required.  Crop the sides if required using the pop-up menu. Equalise the back ground if required using the tool in tab pixel math. Export as stretched JPG or 16 bit bit stretched/unstretched  PNG / TIFF.  The stretched export follows the gamma and stretch setting of the display.  For further image processing you could export to 32 bit float TIFF or 32 bit float PFM format.

ASTAP export types:

File formats ASTAP	8 bit	16 bit	32 bit
Import	FITS, JPEG, PNG, TIFF	FITS, PNG, TIFF, PPM, PGM, raw formats	FITS, PFM
Export	FITS, JPEG, PNG, TIFF	FITS, PNG, TIFF (ASTRO-TIFF),  PPM, PGM	FITS, TIFF, PFM

  


All the program settings and file selections will be save by leaving the program or click on the  Stack check marked images   button.

Back to index



Back to index




The stack menu:



Images, darks, flats can be added using the  Browse   button or can be drag dropped on the form.

For a generic description how to stack see Program operation, stacking astronomical images

Poor images can be filtered by either 1) HFD value,  2) Quality,  3) Star-level or  4) Background value.

Stack multiple objects: Several objects can be stacked in one run. If the classify-by-object is check-marked, the program will stack all objects and save the results using the available darks and flats.

Analysing and removing of the bad ones: Before stacking the images can be analysed with the  Analyse button. Images with a high HFD value are most likely unsharp and can be inspected in the viewer by double click on the filename. The file can be renamed to *.bak by the pop-up menu to be removed later. Images with a high background indicating clouds or twilight should be removed/renamed to *.bak.

Sorting: Images can be sorted on any of the columns. For example if you click on HFD, the images will be sorted on HFD. You could then remove the images with the highest value or inspect them by double click on the file name.

Dark classification: If classify by exposure time and/or exposure time is check-marked the program will automatically select the correct darks or master dark. So it is possible to keep several dark files check-marked in the darks tab and the correct master dark will be selected automatically.

Colour stack and Colour filter classification: If the classify-by-filter is check-marked, the stack routine will combine the available filters to a RGB image. If only Red + Green +Blue image are available they will be combined in a RGB image. If Luminance images are available it will first stack the RGB colors and then apply a most-common-filter and Gaussian blur on the RGB result. Finally the luminance image is coloured with the RGB result. The filter factor should be set typically near 20.

For nebula you could combine RGB or LRGB. If a luminance filter is detected the LRGB mode is used. First the RGB is combined, stars are removed using a most common filter, the image is then blurred with a Gaussian blur and and the colour is then applied luminance channel. Star colour is lost with this process. The filter names can be set in tab alignment

To record star colours use RGB only.

It could be better to stack in two steps. First prepare the Red, Green, Blue and Luminance stacks. Then run a stack again with the Red, Green, Blue and optional Luminance. Or if you could revisit the produced interim results in the RESULTS tab en copy them to the IMAGES tab using the pop-up menu. You could try different colout factors.

Image file names containing "_stacked" will be un-checked by default to prevent stacks by accident are re-used. If required, just select the file and check-makr it again.

Organising images, darks, flats and flat-darks: Images placed in the first tab will be organised based on the FITS header keyword IMAGETYP. So as soon you click on the image  Analyse   button, dark and flats and flat darks/bias images will be move to the corresponding tab.

Keyword modification: The pop-up menu has option to update a keyword of multiple files if required.

OSC images (one shot colour images): To import raw files from a digital camera, ASTAP can either use LibRaw or DCRAW for conversion. You can select it in tab "Stack method". LibRaw has some advantages since the conversion program convert directly to FITS and exposure time, date of exposure and demosaic pattern are written to the FITS header.

The are two option for LibRaw.

• LibRaw (full active area)
• LibRaw  (Cropped active area)

The default value is "LibRaw (full active area). This will extract all active sensor data (e.g. 5202 x3464 pixels) equal A+C below. If you select "LibRaw  (Cropped active area)" you get for a little less area (e.g. 5184 x 3456) equals C below.  

Note that for stacking all images, light, darks, flats, and flat-darks should be of the same dimensions!

The full area M +A+C (e.g 5360 x 3516) could be extracted using the included command-line utility unprocessed_raw using the -F option but has no purpose in ASTAP.

• A + C:   The Active Area, which is the largest area from which a useful image can be formed.
• C:   The Crop Area, which is the subset of the Active Area which many raw converters convert into a useful image. The main reason why C is smaller than A is to provide some extra pixels all around for a raw converter's demosaicing algorithm to use.
• M :  Masked Area, used by some cameras, especially Canons used a dark.

For stacking of  OSC images it is best to start with raw images. The raw colour images look mono, but the program will convert them to colour later in the stacking process. There are four different Bayer patterns. The demosaic pattern can be set in the tab "stack method". Try auto or empirical which will result in the correct colours. A terrestrial image could help find the correct demosaic settings. Load a raw image in the viewer and in tab "Stack method" test conversion with button "Test pattern". Try auto or the four demosaic patterns. If the colours are not correct, just hit undo button or type CTRL-Z to recover and try an other demosaic setting.

Power down option after completion:  If stacking takes a long time you could activate this option and the program will be power-down the computer after completion.

  Clear  , button to remove all files from the list.

    ||      , button to stop blinking cycle.

     ⯈     or      ⯇    , to start a continuous un-aligned blinking cycle. This is intended to find visually outlier images where guiding has briefly failed.

How to exclude poor images

Sort the images on quality by double clicking on the column quality and inspect visually the images with the lowest quality factor by double click on the row. If poor, rename images with right mouse button popup menu "rename to *.bak" for deletion later.  Sort also on background and inspect visually the images with too high values on possible cloud coverage or twilight conditions by double click on the row to open.

• HFD is the image median HFD. The lower the better. Depending on focus and guiding.  Note that low values could be result of loss of tracking. 
• Quality of the image. The higher the better. Based on the number of star detections divided by HFD. Depending on sky transparency of the sky and focus. Note that loss of tracking result both in a low HFD and low star count so a low quality factor..
• Star level. The higher the better. Depending on transparency of the sky and focus. Note that a high values indicate satellite tracks.
• Background. Depending on sky darkness and transparency. The lower the better. A higher value means in most cases a cloud was blocking the sky.
• Sharpness. The lower the better.Measures the change between dark and bright pixels. The measurement is very sensitive to satellite tracks. Could be used to detect satellite track when compared with HFD value. Could also be used to sort images of the Moon and Sun on sharpness.(but they can't be stacked) For correct sharpness measurement of OSC images either the FITS header should contain the keyword BAYERPAT or check-mark "Convert OSC to colour" in tab "Stack method".

To uncheck/untick poor images can als be done automatically. First check mark the option "After analyse untick worst images". Then press button  Analyse and organise images  . The column quality of the images will be analysed statistacally and outliers can be removed using either a  standard deviation  represented by the Greek lower case sigma σ letter or a percentage. 

For a normal distribution you could expect the following:

Confidence interval	Proportion within
1σ	68%
1.5σ	87%
2σ	95 %
2.5σ	98.8%
3σ	99.7%

Satellite tracks

The stack method "sigma clip average" should normally remove any satellite tracks. If after stacking with "sigma clip average" there are still satellite tracks visible, you could lower in tab "Stack Method". the sigma factor from 2.5 to a lower value maybe 2.2 . An other way is the blink/scroll  through the images with the   >>   button. As soon you see an abnormal bright track on the image stop the blinking by esc and inspect visually the image(s)  involved by double click on the row. Remove any poor image by using right mouse button popup menu "rename to *.bak"

Results tab.

The stack results are reported in the results tab. By a double click they can be viewed the viewer. The number of files and exposure times are given. With the pop-up menu it is possible to copy the image file path to the clipboard for use in a file explorer.
 

Back to index

Stack method tab

The best stack option is "Sigma clip average". For only 2 or 3 images or when you are in a hurry or for testing "average"will do.
    

Stack method	Stacking	Description
Average Stacking	✔	For fast stacking. Satellite tracks will not be removed.
Sigma clip average	✔	Stacking, satellite tracks will be removed. Reduce the σ factor for more aggressive filtering of the satellite tracks.
Astrometric image stitching mode	Mosaic	This will stitch astrometric tiles. Prior to this stack the images to tiles and check for clean edges. If not use the "Crop each image function". For flat background apply artificial flat in tab pixel math 1 in advance if required.  Adapt the mosaic canvas height and width if required, default is 2.
Calibration and alignment of the files only		Darks and flats will be applied. The images will be aligned to the reference image.
Calibration of the files only		Darks and flats will be applied.
Average stacking, skip LRGB  combine	✔	Satellite tracks will not be removed. Stacks based on filter will not be combined to RGB.
Sigma clip average, skip LRGB combine	✔	Satellite tracks will be removed. Reduce the σ factor for more aggressive filtering of satellite tracks. Stacks based on filter will not be combined to RGB.

There are two modes of stacking:

• 1) Stacking of image of a grayscale camera or raw images of a DSLR camera.  Stacking in either grayscale of colour goes fully automatic. The program will detect the type of images. Stacking in colour can be forced if the BAYERPAT keyword is not found in the FITS header.
• 2) Stacking of (L)RGB images made with seperate filters. This mode is activated if option Classify by "Light filter" is checked.

Stack method tab,  stacking grayscale images:

There are no special settings for grayscale images. Classify on "Light images" should be unchecked.



Stack method tab,  stacking raw one shot colour images (OSC):  

Classify on "Light images" should be unchecked.
RAW images of  DSLR cameras /One shot color cameras are monochrome and have to be converted into colour images (after applying darks and flats). This converson is called demosaic or debayer. First set Bayer pattern correctly by loading a raw image (grayscale) in the viewer and try one of the bayer patterns till the image colours match in viewer. If not hit CNTRL-Z to undo and try a different Bayer pattern.

There are several methods to convert (demosaic/debayer ) the raw image  to colour:

• In general de-mosaiced OSC astro images are suffering from colour artifacts due to the small size of stars and pixel saturation. If the pixels iluminated by a star are saturated, the red, green and blue values will have the same maximum value and the star centre will appear white. In most case this can be avoid by taking short exposures of 60 seconds or shorter.
• Best results are achieved with de-mosaic methods AstroC and Simple. 
• In most cases the option "Auto level and colour smooth" is required for the correct colour balance and colour smooth. First the three colour channels are adjusted to make the background colour neutral and the stars average white. Secondly the bright stars are smoothed. Both actions can be done manual in the tab pixel math, option "colour correction" and "smart colur smoothing"
  
• If the images are under-sampled and the star colour is random after stacking, use AstroM, white stars. Stars will be whiter. Star colour will be lost. 

The principle of the AstroSimple demosaic method:


Auto levels:  This is an option to white balance the final colour result. The stars will be average white and the background sky will be gray.

Normalise OSC flat:  This option should normally be switched off. Only if your light source for making the flats was very reddish or blueish, you could this option equalise the red, green and blue level. Binning is not recommended for flats since individual sensitive  differences between pixels are compensated by the flat.

Colour smooth:  This is an option to smooth the de-mosaic artifacts. The colour is smoothed while preserving the luminance signal.

Raw conversion.  The program used to convert the RAW file to FITS. It is described here



Oversize: This is default set at zero but could be set hunderd pixel or larger. This overlap was introduced to show partly overlapping images. If you don't want a black area around the image set this value to zero.

If your using a monochrome filter like H-alpha with a DSLR/OSC sensro, use the viewer Tool menu to split the images in R, G, G, B before stacking and use the R=red image for future processing.

The program settings will be saved automatically if your either exit the program or start a stack.



Stack method tab, (L)RGB stacking:

Activating (L)RGB mode:


The filter names should match the filter names in the FITS headers. This is the case if in tab Lights the filter column is displaying red, green, blue or gray icon. If there is no match the icon will be a question  mark.



Stack method tab, Image stiching method (Mosaic)

Astrometric image stiching is possible with the internal astrometric solver. The reference of each pixel is the astronomical position. So stacking is not done against a reference image but against an position array set by the first image.  You have to set the in tab alignment the settings for "Image stiching (method)" correctly. If you stitch 4 images, you have set "Mosaic width/height  in tiles:" at 2. This will provide enough space to place for the 2x2 mosaic.

Here a suggested work method:

1. Stack the tiles separately using method "SIGMA-CLIP-average" and use for the alignment the internal STAR alignment method.  Inspect the resulting tiles and crop them if required. You can also crop them later automatically with "Mosaic skip outside pixels" Do this for each color separately if you have separate files.
2. In tab  "stack method" select option "IMAGE STICHING METHOD" and select astrometic alignment using either the internal solver.
3. In tab "stack method", set the "mosaic width/height" correct and check-mark the option "equalise background".  If the input images have poor borders, set option crop images larger then 0%.
4. Select the files. Most likely the files names contain "_stacked, so you have the check-mark the files after selection.
   
5. Click on the button   Stack check marked images|  
6. Crop the stacked result using the image crop option in the viewer mouse pop-up menu.
7. Adjusted the stretch range and save as JPEG, 90% quality.

Here an example mosaic x 4 of M31 made with ASTAP:





Here an example of a mosaic build of DSS images:

The size can be reduced by a crop function (right mouse button) later. Making the oversize too large could result in memory overload.



Back to index


The alignment menu tab:

For alignment there are four options, internal star alignment,  native astrometric solver, manual alignment or ephemeris alignment. For mosaic building you have to use the internal astrometric solver.



 Internal star alignment



This internal star matching alignment is the best and fastest option to stack images. It is not suitable for mosaics. No settings, fully automatics alignment for shift in x, y, flipped or any rotation using the stars in the image. It will work for images of different size/camera's with some limitations.

The program combines four close stars into an irregular 2D tetrahedron or kite like figure (and compares the six irregular tetrahedron dimensions with irregular tetrahedrons of the first/reference image. It selects at least the six best matches and uses the centre position of the irregular tetrahedrons in a least square fitting routine for alignment.  The four stars are called a quad. The six distances are used to construct a hash code.

There is only two settings relevant but normally you don't have to change them.

• Hash code tolerance  No need to change it. Leave this at 0.007 unless you have severe optical distortion. If you have false detections, set this lower at 0.003.
• Maximum number of stars. Typically set at 500. In some cases like for images with a lot of hot pixels you could set it lower to avoid false detections.

The following image shows the selected quads where the six distances form an irregular tetrahedrons :


The matching process is described here Background info, how does the ASTAP astrometric solving works internally


Astrometric alignment.

   Internal astrometric solver (plate solver). The works with the same four star quad as for the Star alignment option. The quads are compared with the G17 star database (to be installed in the program directory). It has the following settings:

• Field height:  This is the square size of the star field in degrees used for detection and will be set automatically for most FITS files. It should equal to the image height in degrees. If unknown, you could try initaly the option "auto".
• Radius search: Search radius in degrees. If there is no match, the program will move the search field around in a square spiral and increasing the distance form the initial position up to the radius specified.  A radius of 30 degrees could be searched in a few minutes.
• Maximum number of stars to use:  No need to change it. This could be set between 100 and 1000.
• Hash code tolerance:  No need to change it. Leave this at 0.007 unless you have severe optical distortion. If you have false detections, set this lower at 0.003. If you have severe optical distortion or a low resolution image set this at 0.007
• Downsample:  For large image above 3000 pixels wide down sampling will speed up the solving and increase the signal noise ration of the stars. Also colours are combined to monochromatic so this option is beneficial for DSLR images
• Cropping:  For very large image above typical 3 degrees height the solver will only look to the center of the image.
• Force small search steps. This will force a 50% overlap between search fields. Use this if  solving fails. I applied this will slow down blind solvingC 
• Calibrate prior to solving.  If your image is full of hot pixels and soling is less reliable you could try this option. The image will be calibrated with a dark prior to solving removing the hot pixels. Select in the darks tab, a dark or darks with similar exposure duration as the lights you trying to solve. If your select several darks with  different exposure length, check-mark the option classify, "Dark exposure time". 

The internal plate solver works best with raw unstretched and sharp images of sufficient resolution where stars can be very faint. Exposures 5 to 300 seconds. Heavily stretched or photo shopped images are problematic. 

For those are interested:  Background info, how does the ASTAP astrometric solving works internally




Manual alignment.

Manual alignment

This option allows alignment of the images based on a single star, asteroid or comet. If this option is activated, the list of images in the image tab turns red. Double click on each image in the list and click on the star/comet of asteroid to be used as reference. This object is then marked with a little purple circle. The position will be auto centered. (and the X,Y position will be added to the list) A poor lock is indicated by a square. If so try again till it is a circle. If all images in the list are turned green, so contain a value, then click on the   Stack button  .

Options:


For objects which are moving in the sky, select the stack option "average" and not option "sigma clip".

For manual alignment there is a option in the popup menu of tab light to select the next alignment stars automatically:

• Select an alignment star in the first image.
• Select all images in tab lights.
• Select in the popup menu of tab lights "Auto alignment star for selected images"
• If it stops halfway, then is doesn't lock on that image star halfway. Select manually in that image the same star and try again or continue with next images.

Ephemeris alignment

Rather then manual selecting the reference point it is also possible to use the ephemeris of an asteroid or comet. To align by ephemerides go through the follow steps:

Preperation:   

• Configure and test the viewer asteroid annotation menu shortcut CTRL+Q

Stacking:

• Select ephemeris alignment
• Browse for all the images in the image tab and add if available the dark, flats.
• Press the button   Analyse the selected file in tab images  .
• Select from the combobox the asteroid or comet to align on. See screenshot below.
• Hit the  Stack check marked images   button. 
• After the stacking is finished, it is possible to annotate the result.  

Only the solar object selected will be sharp. The stars will form trails

For maximum accuracy it is better to first to calibrate the images with darks,  flats & flat darks. This can be done using  the "calibration only"option in tab "stack method" and then executing the regular stack procedure.

Note that the measured star flux is compared and calibrated with the star database. For most cases you should install the V17 star database containing the Johnson-V magnitudes. After stopping the cycle it is possible to measure any object using the mouse pointer.


Alignment of the images is done using the astrometric (plate) solution. The astrometric solution is written to the orginal file header. You can refresh the photometric and astrometric calibration using the dedicated buttons for this.

The list contains three dates:

• Date/time (start) in YYYY-MM-DDTHH:MM:SS. This is date & time from the header key word DATE-OBS. This is normally the universal time at the start of the exposure.
• JD (mid exposure) This is the Julian Day of the exposure midpoint. This time is calculated from the key word DATE-OBS and half of the exposure time is added.
• HJD (helio) Heliocentric Julian Day at the exposure midpoint expressed in UTC. This is the event time as seen from the Sun center compensation the maximum ± 500 seconds time difference depending on the positions of the Earth, the object outside the Solar system and Sun. For plotting purposes only the fraction is displayed.

To convert the Julian Day to a date and time in the spreadsheet, subtract the Julian Day by -2415018.5 and format as date or time.

For a series of images, the standard deviation of the measured star magnitudes is typical better then 0.02 magnitudes. The standard deviation of the Check star is used for error estimate if more then 4 images are selected. else an estimate based on the Variable SNR values is used. The star flux values should be below saturation (65500)  but reasonable high.

Note that it is beneficial to de-focus an image a little to prevent pixel saturation and spread the flux measurement over more pixels. It also allows longer exposure times. However the image should reasonable focused to allow solving.

Inspector tab.

This tab is intended to measure accurately the tilt and curvature of your telescope & camera setup. It's done by calculating the best focus position for each image area. To do this it requires a series of images taken at several focuser positions.The routine will calculate from these images the best focus point of each area. It will measure the median HFD values of each image and area and build a table HFD as function of the focuser position. Form this data, curve fitting will give the transfer fucntion and the best focus position expressed in focuser steps. The focus point differences between the image areas will indicate the tilt of each area. The difference between the centrum and outer areas focus point will indicate the curvature.

The usage is as follows:

• Prepare a series of short exposure images with different focuser positions and a lot of stars. Exposure time a few seconds. Move for each image the focuser a small fixed step but only in one way to prevent backlash problems. Images with stars having an hfd above 20 will not be analysed correctly.
• Browse with ASTAP inspector tab to the images.
• Press curve fitting.
• The difference between the focus point of each image area will be reported in focuser steps.

The reported hfd values can be selected and copied for further analysis in a spreadsheet.


The image areas "HFD center" 2B (purple) and "HFD out" (any star at more then 75% from center) are as followed defined:




In addition the HFD values of the other eight areas are reported.



Back to index

Live stacking tab.

All file(s) in the specified directory will be stacked live. If it is finished it will wait for new file(s). If a file is detected which is 0.2° away from the previous files a new stack will be started automatically. You can save the stack results from the viewer menu .

To identify files which are processed , they are renamed to the extension *.fts. You can rename them back with the button at the bottom.



Note there is no rejection of poor images. All images are added with equal influence:

Assuming the images are A,B,C,D, E... then

Simple serial stacking:

result1:=A
result2:=(result1+B )/2
result3:=(result2*2+C)/3
result4:=(result3*3+D)/4
result5:=(result4*4+E)/5


Back to index

The reported noise indicated by letter σ is in ADU. The value can be reported in electrons (e-) by setting the option in the status bar popup menu. See other popup menus.


Viewer,  reported angle

If the image the image is solved then the orientation will be reported by a north-east indicator. The reported angle is between the north arrow and the Y axis counterclockwise (rotate east from north) except when the image is flipped. For flipped images it is clockwise (rotate east of north).  So for both flipped and unflipped images it is the angle reported between Y-axis and north arrow in the direction north to east.

A flipped image doesn't represent nature. The flipping (either horizontal or vertical) is caused by the camera system and will change the position of the north arrow. Nature has not changed so to report the correct angle the image has to be unflipped first. You can either unflip it vertical or horizontal but ASTAP is standard unflipping horizontal.

The same angle is reported in the header as keyword CROTA2.  So the angle for flipped images is reported as if the image is "unflipped horizontal".

Header values:



Note about the header: Old style solution keyword CDELT2 is always kept positive and if not the solution is flipped by negating both CDELT2, CDELT2 and shifting the angle 180 degrees. So if the image is flipped the solution is reporting "flipped horizontal" and not an equivalent "flipped vertical". The old style solution is in principle replaced by CD keywords.

Rotation of camera is reported as follows:





Back to index

Viewer,  file menu

If the program is associated with FITS image files or any other format, it will show the image as soon you click image file. Only one instance of ASTAP will be allowed. After clicking on the second image it will be show in the first instance of ASTAP. If you want to open a second instance, just start ASTAP. If the program is started without parameters, you can open more instances.

Images can be loaded from the file menu or can be drag dropped on the main form.

Besides all FITS formats, the viewer support most image formats in 8/24 bit of 16/48 bit format. It can export to any FITS format and 16/48 bit PNG and TIFF formats.
ASTAP can display images and tables of MEF, multi-extension FITS. The images of MEF can be saved as a single image. The MEF tables can copied into the clipboard and paste to a spreadsheet. (v0.9.446)


This ASTAP version can import raw images from almost any digital camera, For this ASTAP executes a modified Libraw tool  "unprocessed_raw" which is included with the ASTAP for most editions. This special version export directly to FITS including date&time, Bayer pattern and active areas of the sensor only.  If "unprocessed_raw" is not included it can be installed in Linux by the "sudo apt-get install libraw-bin" command. 

File formats ASTAP	8 bit	16 bit	32 bit
Import	FITS, JPEG, PNG, TIFF	FITS, PNG, TIFF, PPM, PGM, raw formats	FITS, PFM
Export	FITS, JPEG, PNG, TIFF	FITS, PNG, TIFF (ASTRO_TIFF), PPM, PGM	FITS, TIFF, PFM

The viewer has a preview function. After opening select "Preview FITS files". The preview is displayed in the ASTAP viewer. Use the arrow key to move up or down or just click on the image. The current zoom and position is maintained so you could study the corner of a series images on image quality.

The file open menu with preview selected:


Back to index



ASTAP viewer screen shot:




Viewer menu Tools:


Viewer, tools, Astrometric solving the image in the viewer

  
With the   Solve|   button it is possible to find an astrometric solution of the image loaded. For this the estimated celestial center position α, δ should be available. This position is normally retrieved from the FITS header. Secondly the estimated image height in degrees should be specified in the stack menu, tab alignment.  In the same tab alignment you can specify the search radius and down sampling. For successful solving see conditions required for solving.

For solving JPG, PNG or RAW files it is possible to add the object position as center position using the deep sky database. Double click on the δ position in viewer and enter the object name. The position will be retrieved from the database. This position will be used as a start point for the solver.



Back to index


Add SIP coefficients to the header

This option will add basic SIP polynome coefficients to the header to cope with basic barrel or pincushion distortion. In addition the image scale is adapted for the center undisturbed part only. The scale in the outer regions could be a fraction of a percentage different from the center of the image due to the distortion. Normally the scale factor is a compromise between the center and the slight disturbed outer regions.

The SIP correction can be tested with the option "Show distortion"

Back to index

Viewer, Tools, Batch processing:

With the batch routine several FITS image can be "astrometric solved" or converted. This conversion is not required for ASTAP. Automatic conversion is integrated in the menus.

Back to index


CCD inspector

The viewer has a CCD inspector under the menu tools.  By reporting the median HFD value, sensor tilt and curvature the inspector will quickly show any problem of the imaging system. 

The image used for testing should be a single raw image with sufficient stars and not containing a disturbing large and bright deep sky object.  The exposure length should be long enough to image many stars but not too long preventing star saturation.

The routine will detect and annotate the stars with their HFD value and plot a tilt indicator in the image. A star rich image containing a few hunderd stars or more gives the best accuracy.

The following will be reported:

• Image median hfd which is an excellent indicator of the quality of focusing. The lower the value the better the focusing, the sharper the stars are. The value is also depending on the astronomical seeing and the quality of the optics. If the image is solved it will also be indicated in arc seconds.
• Sensor tilt as the HFD difference between the best and worst corner median values. In addition as an graphical indication it draws an trapezium in the image based on the four median values.
  There can be some variation in images of the same series. Any tilt equal or less then 10% of the range is a good value. Any tilt equal or more then 20% of the range indicates a tilt problem.  
• Off-axis abberation as the delta between the HFD value in the center and in the outer areas of the image. The stars in the outer area are normally a little larger, oval or comet shaped due to the optics or curvature giving an higher HFD value. The lower this value the better the optics. This value could  be a help to adjust the optimal distance between the field flattener and the camera. This measurement is only a valid if the focus for the center area is perfect. A more advanced measuring method is available in tab CCD Inspector.

An irregular octagon figure is displayed using the median HFD value of eight areas. If your optics is good and there is hardly any tilt or curvature the figure will form a square:. The yellow values indicate the median values of nine square areas:

The figure of a good optical system:



A system suffering from severe curvature:


A system with severe tilt:


It is possible to automate the loading and tilt measurement using live stacking


CCD inspector calculation method explained:



The tilt is calculated as follows:







Triangle: The triangle based tilt indication is intended for adjusting your tilt adapter with three adjustment screws. First solve an image, then flip the image in the viewer such that north is up and east is left (northern hemisphere). Then apply the adjustable three corner angle such that the orientation of the adapter screw matches with the three corners show. You can click on the tilt button without refresh.

The triangle based tilt indication is based on the following three areas:




The triangle option measures in an circular area with a diameter equals the image height. This area is split in three equal 120 degrees segments. The center is excluded.

It is possible to automate the loading and tilt measurement using live stacking

For the triangle the off-axis aberration calculation is as follows:


The tilt calculation:



The diameters of the areas 1,2 and 3 are equal to the distance to the longest side of the image. The triangle option is less sensitive then the octagon method but the areas are symmetrical so the orientation can be adapted without changing the sensitivity.

Normally the tilt indication uses stars with a signal to noise ratio (SNR) of 30 or higher. For extra stars is will reduce the minimum SNR to 10. Then the HFD values will be a little less accurate. Extra stars is intended for short exposure only where there are not enough stars with a SNR of 30 or higher.

HFD 2D contour.

The star half flux diameters (HFD's) are displayed in a 2D contour. Dark areas indicate a lower and better HFD value. This will quickly visualise sensor tilt or other problems. To avoid false indications by outliers the HFD values are filtered by taking the median HFD value of the three closest stars and  allocate the result to all three stars. The HFD values are indicated numerical. They grey levels have no direct linea relation with the HFD.




HFD diagram.

The star HFD values can also be represented by areas of constant HFD. It is in principle a Voronoi diagram, but by  taking the median value of each three closest stars and allocate the median to the three stars it looks a little different.White areas indicate a star with an high HFD value. Darker areas indicate a lower value. The grey level is the HFD * 100.


HFD values

This tool will only indicate the median filtered HFD values next to the stars. The same values as in the 2D countour and HFD diagram. To avoid false indications by outliers the HFD values are filtered by taking the median HFD value of the three closest stars and  allocate the result to all three stars.

Unroundness

This tool measured the unroundness of the imaged stars. The values are the aspect ratio of an ellipse.

Measuring principle: The star is split in two by a line. The average distance of the pixels to the split is measured. Then the split line is rotated one degrees and again the average pixel distance to the split line is measured. This continues till the line has made a 180 degrees rotation.  The aspect ratio is the highest distance value found divided by the lowest distant value. The orientation is the position where the lowest distance is found en the star is the longest.

  This unroundness measuring principle is licensed under a Creative Commons Attribution 4.0 International License. 

Median background values

This tools writes the median background values as numerical values in the image. Stars will be ignored but nebula will influence the background measurement..





Show distortion

This tools shows the difference between the Gaia star positions and the centroids of the imaged stars assuming a linear solution. A difference is indicated with a line 50 x larger then the actual difference in pixels. A scale is show at the left bottom. Also the  median astrometric error in arc seconds for the center 50% square (height/2 * height/2) is reported. This indicates the error to expect for astrometric measurements. Note the database resolution is 0.077"  in α and 0.039" in δ.

If SIP polynome as measuring mode is selected then the distortion will be partly corrected. The SIP polynome coefficients can be added by the tools option "Add SIP coefficients to the header". This correction will only work for basic barrel or pincushion distortion. You could also use the advanced SIP solution from Astrometry.net but select "order" three rather then the default two.

Back to index


Aberration Inspector

This tool creates a 3x3 mosaic of the images center, the corners and borders. This allows an easy comparison of the star shape at the different sections of the image.



Back to index



Viewer, Tools, Calibrate Photometry.

With this option the relation between flux and magnitude is measured. The image should be solved in advance to be able to calibrate the star flux with star database magnitudes. After the calibration is applied, the star magnitude at the mouse cursor is displayed. In the log an estimate for the limiting magnitude for point sources is reported for a detection limit of SNR ≥7.  The accuracy is better then ±,0.5 magnitudes. The aperture used for star flux measurement can be adapted in stack menu tab "photometry".

For stretched images the reported (limiting) magnitude will be less reliable. Use the popup menu option "online query" to request the magnitude of the point sources.



Viewer, Tools, SQM report based on an image



With this option the  SQM = sky background value relative to the stars is measured and expressed in magnitude per square arc second.  The image should be solved in advance to be able to calibrate the measured star flux with star database magnitudes. The reported SQM value will be equal to a value reported by an Unihedron SQM-L meter.  Atmospheric extinction of the stars at lower altitudes will be compensated.

Some background information:

http://www.lightpollution.it/download/sqmreport.pdf

At zenith the measured star flux and sky background flux are defined as comparable and star light extinction is zeroed by subtracting 0.28 magnitudes from the calculated zenith extinction. So at zenith the SQM brightness values are comparable to deepsky object brightness and can be expressed in magn/arcsec². At lower altitudes the measured star flux is less and compensated by a predicted extinction (0.2811*airmass, airmass according Pickering).

Pre-conditions
1) Image is astrometrical solved for star flux-calibration against the star database magnitudes.
2) The background value is larger then pedestal value or mean dark value. If not expose longer.
3) Apply on single unprocessed raw images only.
4) Providing dark image(s) in tab darks (ctrl+A) or entering a pedestal value  (mean value of a dark)
     increases the accuracy. If possible provide also a flat(s) in tab flats.
5) DSLR/OSC raw images require 2x2 binning. For DSLR images this is done automatically.
6) Most of the image is free of deepsky nebula.
7) The calculated altitude is correct. The altitude will be used for an atmospheric
    extinction correction of the star light.
8) The H17 or H18 star database is used containing the Gaia blue magnitudes (400-700 nm) and no filter is used except a UV/IR block filter.  


Differences between Unihedron and ASTAP measurements:

• The Unihedron is an absolute measurement of the sky glow. The ASTAP measurement is a relative to star light measurement. In case of poor sky transparency ASTAP will report lower SQM values due to the higher extinction of star light.
• It is expected that the ASTAP measurement will be less sensitive to Milky Way light since it measured the background and not the total sky glow.
• The Unihedron measurement will be less sensitive to the blue part of the spectrum.

Viewer, Tools, Magnitude (measured) annotation.

With this option the stars are annotated with the estimated magnitude.  The image should be solved in advance to be able to calibrate the star flux with star database magnitudes.

If the Johnson-V version of the star database (V17) is used, the results match very accurate with AAVSO charts as demonstrated below. Camera was an ASI1600 with only an UV-IR block filter:



For best accuracy the image should be monochrome and the Gaia Johnson-V star databases V17 should have been installed and selected. The image should have taken with a Johnson-V filter or none (clear). Saturated stars will be ignored since it is not possible to measure then accurately.



In the left bottom corner of the image an estimate for the limiting magnitude for point sources is reported using a detection limit SNR ≥ 7. Below this value detection of point source detection becomes unreliable. The accuracy is better then ±,0.5 magnitudes. The image should not be stretched. The aperture used for star flux measurement can be adapted in stack menu tab "photometry". Results can be validated by requesting the Gaia BP magnitude magnitude of the faintest stars in the image using the popup menu option "online query".

For stretched images the reported limiting magnitude will be less reliable. Use the popup menu option "online query" to request the magnitude of the faintest sources.


Back to index


Viewer, Tools, Star (database) magnitude.

With this option the stars are annotated with the star database magnitude. This can be done best with V17 containing the calculated Johnson-V magnitudes.

For a G-database the indicated magnitude is Gaia blue. For a V-database the indicated magnitude is Johnson-V  and the following the difference between Gaia blue and red,  positive for reddish objects. All in 1/10 of a magnitude.

Below, the image is 1) solved, 2) auto calibrated (using the V16)   The cursor is at a star and based on the flux of all know stars, the star Johnson-V magnitude is estimated to be 16.1. The stars are marked with the Johnson-V magnitude and  Bp-Rp color indication.

See also blink and photometry




Back to index

Viewer, Tools, Unknown star magnitude

Any unknown object or a star with an abnormal magnitude is identified. This option is intended to mark nova and minor planets using the star database. Any star like object missing from the star database will be marked. Also if the measured magnitude is one magnitude brighter then the database. The H18 star database up to magnitude 18 is recommended for this option. If the H18 is selected then missing object up to magnitude 17 will be detected. The image should be solved in advance.


Viewer, Tools, Variable star annotation

Variable star are annotated using the variable_stars.csv database.



Viewer, Tools, Asteroid and comet annotation

This option will annotate asteroids using the orbital elements taken from the MPCORB.DAT file and for comets the ComeEls.txt from the Minor Planet Center.

Usage:

- Solve an image.
- Go to to the viewer "Tools" menu, "Asteroid annotation".
- First time download the full MPCORB.DAT from the minor planet center. Link is available from the blue down arrow. Set the path to MPCORB.DAT correct.
- Set the limiting magnitude and maximum number of asteroids to read.
- Press the button   Asteroid & comet annotation   .

Remarks:

Renew the MPCORB.DAT and CometEls.txt  files every few months.

The observation date and time are extracted from the FITS header (date-obs, date-avg) or for other files the file date is taken.  If date average is not available it will be calculated from the exposure time and date-obs from the FITS header.

The latitude and longitude of the observation location are also taken from the header. If not available enter them manually or leave them at 0.


Back to index

Viewer, Tools, Deep sky annotations  

If the image is solved, it is possible to add deep sky annotations. See pull-down menu TOOLS:



Back to index


View menu

This menu has the following options:

• Image clean up Ctrl+space
• Center lost windows Ctrl+F12  Use this if you have multiple screens and once window is out of site for some reasons.
• Flip horizontal Ctrl+H
• Flip vertical Ctrl+H
• Zoom out  PgUp
• Zoom in     PgDn

Back to index

Viewer popup menu:



Add annotation,  free text label at a x,y position. It can be connected via a line by first holding the right mouse button and moving the mouse away. See sample above.  Persistent by annotation keyword in the FITS header. The annotation can be switched off in pulldown menu "View". Remove by removing the annotation line in the header.  If a @ is added to the text, the annotation is written persistent in the image data. So the image will be permanently altered after saving. Adding two or more @ will increase the font size.

Add marker, rectangle marker at x,y position. Draw the rectangle first by holding the right mouse button down and moving the mouse away. See sample above. The Persistent by annotation keyword in the FITS header. The annotation can be switched off in pulldown menu "View". Remove by removing the annotation line in the header.

Add object position, Enabled after astrometric solve. Will add the α, δ at the mouse position. Orange if a lock is possible, see above sample, red if not. Not persistent after image clean up.

Add marker at  α,  δ position, will place a yellow square at the specified α,  δ position.  See above sample for orange star. Enabled after astrometric solve. Persistent. Position will be saved with "save settings". If a C is typed, the marker is placed at the center.

Measure total magnitude within rectangle, Enabled after astrometric solve. The program will measure the flux and try to estimate the magnitude. Hold the right mouse down and pull a rectangle around the deep sky object, release mouse button and then select this menu.  For best accuracy the image should be monochrome and one of the Johnson-V star databases should have be installed and selected (V16 or V17). The image could have been taken with a Johnson-V filter for best results or none. The measured pixel values should be below saturation.



A demonstration is available at YouTube:  Photometry in the viewer

The measuring principle is as follows:

1. Use the star database to measure the MEDIAN relation between flux of the detected stars and star magnitude from the database. (That's why solving is required and best result are achieved with the V16 or V17 Johnson-V star databases based on Gaia DR2) 
2. Measure the MEDIAN background 1 to 10 pixels wide outside the rectangle box. This median measurement will ignore stars in the field.
3. Measure the MEAN  flux inside the box.
4. Calculate the magnitude for the "inside mean flux" minus "median outside box flux" using the relation found for the stars.

 You could argue that a Johnson-V filter or green channel is required for the image but in practice the error is limited depending on the spectrum.

Remove all markers and labels. Will remove all none persistent labels.  Persistent annotations can be switches off in the viewer pulldown menu VIEW.

Copy image (selection) to the clipboard, copies the displayed image or a selection of the image to the clipboard. The orientation is depending on the selection direction.. (v0.9.417)
Copy position to clipboard, Enabled after astrometric solve. Copies the α,  δ position to the clipboard.
Copy position to clipboard in ° , Enabled after astrometric solve. Copies the α,  δ position in degrees to the clipboard.



Online query. Query the Simbad database, Hyperleda, Ned or Gaia online database for all objects within the selected rectangle. Or request an AAVSO map for the selected rectangle.  The annotation requests will annotate the objects in the image. The browser requests will activate the default web browser and list the object found in a table with information and further links. For a Gaia query it is better to place the cursor directly on the star rather then pull a rectangle.

Local adjustments

Gradient removal tool

To remove a linear gradient caused by light pollution or twilight. The routine need two empty areas 40x40 pixels in the image to measure the gradient. The area may contain stars but no deepsky object. The area are selected by pressing the right mouse button(first area) and while holding the mouse button move to the second area in the direction of the gradient. Then select in the popup menu "Gradient removal tool". Try to maximise the distance ideally the full image range.




Copy paste tool.   Powerful touch-up tool  for cosmetic corrections. Small sections of the image can be copied as pasted on hot pixels or artifacts. Select the good part by holding the right mouse button and pulling a rectangle with the mouse and move the copied part to the part to be touched up..


Set area.  Set an area for the colour replacement tool in tab pixel math 1.


Remove deep sky  object. Removes an deep sky object as part of the pixel math tab routine "equalise background tool".  Hold the right mouse down and pull a rectangle around the deep sky object, release mouse button and then and then select this menu.


Remove borders. This menu allows to remove parts of the image near de borders.


Crop fits image. This allows the crop the image. Hold the right mouse down and pull a rectangle, release mouse button and then and the select this menu

Back to index

Other popup menus:

Search for an object position:

Copy histogram values to clipboard. You can paste them then in a spreadsheet.


Fits header editor



Status bar. You can select to display:
 
- A different unit for the position.
- The HFD and FWHM in arc seconds.
- Noise reported in electrons.



Back to index

---

Conditions required for solving:

Quick checklist for solving:
If you still have problems with solving, you could send me a link to the file involved (FITS is preferred) and if I have time I could have a look. Upload it e.g. to http://nova.astrometry.net or https://ufile.io/ and give me the link. For privacy reasons, prior to uploading you could remove your latitude/longitude information from the header using the viewer "Tools", "Batch processing", "Remove longitude latitude information" menu.

The internal astrometric solver works best with raw unstretched and sharp images of sufficient resolution where stars can be very faint. Heavily stretched, saturated, out-of-focus or photo shopped images are problematic. It requires minimum about 30 stars in the image to solve. Images containing of a few hundred stars stars are ideal. For star rich images, the program will reduce the detection limit to limit the number of stars. This will only work for unstretched images where brighter stars have a greater intensity then fainter stars. So ASTAP requires three star dimensions for solving. The star x, y coordinates and star intensity. Oval stars due to tracking errors or severe optical distortion will be ignored and solving could fail.

Check list for successful solving with ASTAP:

1. An approximate α and δ celestial position is specified (for a spiral search). For FITS file this is normally read from the FITS file header and set automatically. The position can also be passed by the command-line. In case you view an JPEG, TIFF image, double click on α input to search for a know deep sky object position from the database.
2. Correct image height in degrees specified within preferable 5% accuracy.  See ∑ window, tab alignment, group-box astrometric settings, "Field of view (height)". For FITS images this is normally automatic calculated from info in the file header. In cause of doubt you could try "Field of view (height)"=AUTO
3. At least 30 stars are visible in the image. They can be very faint, barely visible in the noise. For large field-of-view (>1°) expose 5 to 10 seconds but for small field-of-view (<0.5°) far away from the Milky-Way plane you have to exposure longer up to 60 seconds and use the H18 star database.
4. Reasonable round stars and the camera in focus. You could verify the star detection by the CCD inspector or by  the  Test button to show quads  in the  ∑   window, tab alignment). Most stars should be detected.
5. Image dimensions at least 1280x960 pixels. Set downsample at 0 (=auto) and the program will select a downsampling factor automatically. Image height in pixels after ASTAP dowsampling should be somewhere between 1000 and 3000 pixels. ASI120MC camera image are problemetic to solve.
6. Image is not stretched, very saturated or heavily photo-shopped. The total exposure time could be hours as long it is possible to separate the brightest stars from the faint by intensity.
7. Search radius should be set large enough. See  ∑   window, tab alignment, group-box astrometric settings.  You could set this at 30° or for blind solving at 180°.
8. Use downsample factor 0 (auto) or 2 for raw DSLR/OSC images. See  ∑   window, tab alignment, group-box astrometric settings. Save settings if modified (Viewer pull down menu, file, exit (and save settings). 
9. If your image is full of hot pixels you could adjust the "Ignore stars less then["]"  in tab Alignment. This is set default at 1.5" but could be set higher for a long focal length.
10. An other option for hot pixels is option "calibrate prior to solving". Select in the darks tab, a dark or darks with similar exposure duration as the lights you trying to solve. If your select several darks with  different exposure length, check-mark the option classify, "Dark exposure time". 
11. The maximum number of stars to use should be defined. Typical set at 500. See ∑ window, tab alignment.
12. Hash code tolerance should be defined. Typical set at 0.007. See tab alignment.
13. For some solver failures you could force in ASTAP the option "slow" (-speed)  for more reliable solving. The reliability will be very high but solving will be slower.
14. In case of excellent seeing the imaged star could be smaller then 1.5".  Reduce the setting "Ignore stars less then 1.5" in tab "Alignment". Star size should 2 pixels or more. Adapt downsampling accordingly. Improvements will be reported by a increased number of quad matches.
15. For a field of view less then 60 arc minutes (long focal length) it is recommend to install the H18 star database and remove the H17 (or force use of the H18 in tab alignment). 
16. If a global cluster fills the whole field, ASTAP could struggle to solve. Forcing option "slow" could  help.
17. Best input format is FITS, RAW or 16 bit PNG, TIFF. JPEG compression or 8 bit PNG, TIFF files are a disadvantage.

Test if an image is solvable (full blind solving):

To test if an image is solvable, follow this procedure:

A) Load an image in ASTAP

B) Check star detection by hitting button F4. Only stars should be annotated and not the hot pixels. If hot pixels are annotated see E) Hot pixels seen as stars

C) Go to ASTAP settings by CTRL+A or Σ button. Select tab Alignment. Set ASTAP to blind solving by 1) in screen shot

If too many hot pixels are detected as stars then increase "Ignore stars less then ["]" value marked by 2) in the screen shot. The star annotation values shown after hitting F4 are indicating which value is you should use. More then the hot pixels HFD values but less then the star HFD values. Save settings by hitting the solve button.

 The FOV calculator in tab Pixel math 2

Warnings:

• In case you get regular a warning "Star database limit reached" replace the G17 database by the H18.  This could happen if your field of view is small (<30 arcmin). Remove the G17 files to force the use of the H18.

In the plane of the Milky-way plane it is easy to image enough stars for solving as shown in this ESA image showing the sky star density. If you look outside the Milky-Way plane into deep space there are less stars. The G17 will contain enough stars if your field-of-view is 0.5 degrees or more. For a smaller field-of-view you have to exposure longer to image fainter stars. The H18 could work well up to a field-of-view of 0.25 degrees is exposed enough. For large field-of-view (>1°) expose 5 to 10 seconds but for smal field-of-view (<0.3°) you have to exposure longer 30 to 300 seconds. Expose long enough till the solver detects 30-50 stars minimum or till the solver gives a warning "Star database limit reached" indicating you have reached magnitude 18.


Support is available via the ASTAP Forum

Example of the INI output file after an successful solve:

To analyse a FITS file you could do in a Windows batch file the following:

c:\astap.fpc\astap.exe -f  c:\astap.fpc\test_files\command_line_test\m16.fit -analyse 30
echo Exit Code is %errorlevel%
pause

You will get
Exit Code is 326000666

where the HFD is 3.26 using 666 stars

For Linux and Mac a stdout is used reporting as follows:

HFD_MEDIAN=3.3
STARS=666

Program	Windows	Linux	macOS
astap	exit code	stdout	stdout
astap_cli	exit code & stdout	stdout	stdout

Finding best focus based on four or more input images:

c:\astap.fpc\astap -focus1 D:\temp\FocusSample\FOCUS04689.fit -focus2 D:\temp\FocusSample\FOCUS05039.fit -focus3 D:\temp\FocusSample\FOCUS05389.fit -focus4 D:\temp\FocusSample\FOCUS05739.fit -focus5 D:\temp\FocusSample\FOCUS06089.fit -focus6 D:\temp\FocusSample\FOCUS06439.fit -focus7 D:\temp\FocusSample\FOCUS06789.fit -focus8 D:\temp\FocusSample\FOCUS07139.fit
echo Exit Code is %errorlevel%
pause

or with the -debug option

astap.exe  -debug -focus1 D:\temp\FocusSample\FOCUS04689.fit -focus2 D:\temp\FocusSample\FOCUS05039.fit -focus3 D:\temp\FocusSample\FOCUS05389.fit -focus4 D:\temp\FocusSample\FOCUS05739.fit -focus5 D:\temp\FocusSample\FOCUS06089.fit -focus6 D:\temp\FocusSample\FOCUS06439.fit -focus7 D:\temp\FocusSample\FOCUS06789.fit -focus8 D:\temp\FocusSample\FOCUS07139.fit

Or with the special command line version:

Solving should be reliable. In case of failure, have a look to conditions required for solving.

For NINA version 1.9

Solving should be reliable. In case of failure, have a look to conditions required for solving.

Progress is shown in tray icon and popup notifier.

If you select in SGP the PlateSolve2 setting "Max Regions" (=3000) this will force a search up to 90 degrees diameter. However the solver will be most likely stopped halfway by a SGP timeout.

---

Installation of the external Astrometry.net solver:

MS-Windows:

Install a local copy of Astrometry.net (via ANSVR  or  Astrotortilla)   as the astrometric solver. Or alternatively if you have Win10, 64 bit Creation edition you use the new Linux sub-system

ANSVR: The ANSVR link contains a newer compilation of astrometry.net made for SGP. It runs as a Linux program under Cygwin in MSWindows. Follow up to installation step 9. The link you have to put in ASTAP is as follows:

Appendix 1, the stack process:

Back to index

Appendix 2, Why use flat-darks

The main reason why you have to use flat-darks or bias frames is that they provide a pedestal so a fixed value to the flat routine. The same pedestal value which is present in the flat. This is typical a pixel value of something around 500 to 1000. Without this value the flat compensation will not work properly, and the darker corners of the light frame(s) will be under compensated.

This can be visualised with a simple X, Y diagram. The green part of the value is due to the light exposure. Blue is the pedestal value. If the flat-dark is included (subtracted) the intensity at the corners of the sensor drops to 40000/50000 equals 80%. If no flat-dark is included, then the measured intensity drop will be 41000/51000 equals 80.3%. This will result in an under compensation by the flat routine and the corners of the stack will be a little darker than the center. So use flat-darks or bias frames for best flat correction.

Some cameras behave a little weird in the first second. For those it could be better to use the same exposure time for the flat-dark as for the flat.  A bias is a flat-dark with an exposure time 0 or about zero. For most cameras you will not notice the difference between using either a bias or flat-dark. 

Back to index

Appendix 3, the 1476 and 001 star database format:

The 1476 format:

The .1476 format divides the sky in 1476 area's and 1476 corresponding files with the extension .1476. The 1476 areas have a minimum width or diameter of 90°/17.5 equals 5.143°.  The boundaries are defined by constant α and δ values. For the similar 290 file format see the HNSKY help file

Each star is stored in a record of 5 bytes. The files start 'with a 110 byte header containing a textual description and the record size binary stored in byte 110.  The RA values are stored as a 3 bytes word. The DEC positions are stored as a two's complement (=standard), three bytes integer. The resolution of this three byte storage will be for RA: 360*60*60/((256*256*256)-1) = 0.077 arc seconds. For the DEC value it will be: 90*60*60/((128*256*256)-1) = 0.039 arc seconds.  The stars are sorted on magnitude and the magnitude is stored in a byte of the special preceding header with an offset to make all values positive.

  Example of star Sirius RA and DEC position:
  The RA position is stored as C3 06 48 equals: (195 +6*256 +72*256*256)*24/((256*256*256)-1)=6.75247662 hours equals: 6:45:8.9
  The DEC position is stored as D7 39 E8, equals: 215 57 -24. The DEC is then (215 +57*256 -24*256*256)*90/((128*256*256)-1)=-16.7161401 degrees equals -16d 42 58

Record format:

The 290-5 or 1476-5, short record size of 5 bytes for one star without designation:

    type
      hnskyhdr290 = packed record
        ra7  : byte;
        ra8  : byte;
        ra9  : byte;
        dec7 : byte;
        dec8 : byte;{magnitude and dec9 are written once in preceding header-record}
      end;



Stars are sorted from bright to faint in the "0.1" steps. Within the magnitude range, the stars are additional sorted in DEC. For a series of stars with the same DEC9 value, a header-record is preceding containing the DEC9 value stored at location DEC7. Since the stars are already sorted in 36 declination bands, the number of DEC9 values is already limited by a factor 36.

1476-5 header-record example: FF FF FF 20 06 This indicates the following records have a DEC9 value of 20 -128 offset and a magnitude of (06 - 16)/10 equals -1.0 (new method, +16 offset).

The shorter records methods become only space efficient for very large star collection of a few million stars. In these large collections many stars can be found with the same magnitude and DEC9 shortint. The Gaia database is only issued in the 1476-5 format of 5 bytes per star. or in an older format 290-6 (V17) or 290-5. The 290-6 has one more byte for the colour information. This is documented in the HNSKY planetarium program help file.

So the record sequence will be as follows:


The 1476 areas: Each ring is divided by lines of constant RA such that the minimum width in RA is about 5 degrees (deltaRA*cos(DEC)
.

Declination minimum	Declination maximum	Ring	RA cells	RA step north[degr]	RA step distance south[degr]	DEC step		Files
-90.00000000	-87.42857143	0-1	1					0101.1476
-87.42857143	-82.28571429	1-2	3	5.38377964	16.10799190	-2.57142857	{=90/(2*17.5)}	0201.1476, 0202.1476, 0203.1476
-82.28571429	-77.14285714	2-3	9	5.36933063	8.90083736	-5.14285714	{=90/17.5}	0301.1476,  . . . . . , 0309.1476
-77.14285714	-72.00000000	3-4	15	5.34050241	7.41640786	-5.14285714	{=90/17.5}	0401.1476,  . . . . . , 0415.1476
-72.00000000	-66.85714286	4-5	21	5.29743419	6.73757197	-5.14285714		0501.1476,  . . . . . , 0521.1476
-66.85714286	-61.71428571	5-6	27	5.24033376	6.31824883	-5.14285714		06 . . .
-61.71428571	-56.57142857	6-7	33	5.16947632	6.00978525	-5.14285714		07 . . .
-56.57142857	-51.42857143	7-8	38	5.21902403	5.90674549	-5.14285714		08 . . .
-51.42857143	-46.28571429	8-9	43	5.21991462	5.78564078	-5.14285714		09 . . .
-46.28571429	-41.14285714	9-10	48	5.18296987	5.64803600	-5.14285714		10 . . .
-41.14285714	-36.00000000	10-11	52	5.21357169	5.60088688	-5.14285714		11 . . .
-36.00000000	-30.85714286	11-12	56	5.20082354	5.51859939	-5.14285714		12 . . .
-30.85714286	-25.71428571	12-13	60	5.15069276	5.40581321	-5.14285714		13 . . .
-25.71428571	-20.57142857	13-14	63	5.14839353	5.34991355	-5.14285714		14 . . .
-20.57142857	-15.42857143	14-15	65	5.18530082	5.33887123	-5.14285714		15 . . .
-15.42857143	-10.28571429	15-16	67	5.17950194	5.28678585	-5.14285714		16 . . .
-10.28571429	-5.14285714	16-17	68	5.20903900	5.27280509	-5.14285714		17 . . .
-5.14285714	0.00000000	17-18	69	5.19638762	5.21739130	-5.14285714		18 . . .
0.00000000	5.14285714	18-19	69	5.21739130	5.19638762	-5.14285714		19 . . .
5.14285714	10.28571429	19-20	68	5.27280509	5.20903900	-5.14285714		20 . . .
10.28571429	15.42857143	20-21	67	5.28678585	5.17950194	-5.14285714		21 . . .
15.42857143	20.57142857	21-22	65	5.33887123	5.18530082	-5.14285714		22 . . .
20.57142857	25.71428571	22-23	63	5.34991355	5.14839353	-5.14285714		23 . . .
25.71428571	30.85714286	23-24	60	5.40581321	5.15069276	-5.14285714		24 . . .
30.85714286	36.00000000	24-25	56	5.51859939	5.20082354	-5.14285714		25 . . .
36.00000000	41.14285714	25-26	52	5.60088688	5.21357169	-5.14285714		26 . . .
41.14285714	46.28571429	26-27	48	5.64803600	5.18296987	-5.14285714		27 . . .
46.28571429	51.42857143	27-28	43	5.78564078	5.21991462	-5.14285714		28 . . .
51.42857143	56.57142857	28-29	38	5.90674549	5.21902403	-5.14285714		29 . . .
56.57142857	61.71428571	29-30	33	6.00978525	5.16947632	-5.14285714		30 . . .
61.71428571	66.85714286	30-31	27	6.31824883	5.24033376	-5.14285714		31 . . .
66.85714286	72.00000000	31-32	21	6.73757197	5.29743419	-5.14285714		32 . . .
72.00000000	77.14285714	32-33	15	7.41640786	5.34050241	-5.14285714		33 . . .
77.14285714	82.28571429	33-34	9	8.90083736	5.36933063	-5.14285714		34 . . .
82.28571429	87.42857143	34-35	3	16.10799190	5.38377964	-5.14285714		3501.1476, 3502.1476, 3503.1476
87.42857143	90.00000000	36-37	1					3601.1476

The 1476 areas:


South pole view of the 1476 areas:



The 001 format:

For wide field the stars are stored in a single file w08.001 down to magnitude 8.  The stars are sorted from bright to faint. The file starts with a 4 byte integer specifying the number of records=stars and has for the W08 value 41246. Each star is stored in a record of three singles (4 byte floats). Starting with magnitude [x10] then  RA [radians]  and finally DEC[radians]. So the data is written as the following array (first star is Sirius):

  star_array : array[0..41264,0..2] of single =    //all stars up to magnitude 8
  (
  (  -15    ,    1.767725252    ,   -0.291899504   ),
  (   -6    ,    1.675309906    ,   -0.919709903   ),
  (    0    ,    4.873596766    ,    0.676937966   ),
  (    0    ,    3.837100429    ,   -1.061694783   ),
  (    0    ,    3.73338603     ,    0.334553937   ),
  (    1    ,    1.381830995    ,    0.802764629   ),
  (    3    ,    1.372430608    ,   -0.143145706   ),
  (    4    ,    2.003995701    ,    0.091067934   ),
 

Back to index