Features in the Gallery: Sometimes you can do better

The image from 2009-04-20 01:54:28 shows a typical IRAS 100 micron image:
IRAS 100 micron image
There’s a very detectable pattern of lines crossing the image from top left to bottom right. These are remnants of the scan lines used in taking the data. The team that built the IRIS survey, headed by Marc-Antoine Miville-Deschenes and Guilaine Lagache reprocessed the data to — among other things — get rid of, or at least minimize, these scan lines. If we simply choose the IRIS100 survey with identical parameters we get the image below. The scan lines are completely gone. That’s why it’s IRIS: the ‘Improved’ Reprocessing of the IRAS Survey. Nice job guys!
IRIS 100 image

This entry was posted in Discussion and tagged , . Bookmark the permalink.

13 Responses to Features in the Gallery: Sometimes you can do better

  1. Karl says:

    It appears that IRIS reprocessing is suppose to make IRAS Data more distinct and clearer.

    Can someone explain these IRIS, SFD100um and IRAS SkyView INFRARED Images near equatorial coordinates: 13h49m0330s-8d34m31?‏



    They appear to show less detail in the reprocessed Images than the original IRAS Survey.

  2. Tom McGlynn says:

    I believe these may have been caused by observations of a solar system object — likely Jupiter. It looks like they were cut out of the data for IRIS. Note that the IRIS100, SFD100um, and IRAS100 are all using the same underlying observations.

    Most of the sky was observed more than once by IRAS and I believe images from the individual scans are available from IPAC. The epoch of these images could be compared with the position of Jupiter to see if the objects show up in the expected images.

  3. Tom McGlynn says:

    I did download the HCON images for the specified region from the LAMBDA archive (http://lambda.gsfc.nasa.gov). These are the independent scans. LAMBDA provides both the summed images used in the SkyView surveys and the indivdidual HCONs. A single object shows up on each of HCON1 and HCON2 consistent with the idea that it’s a solar system object in orbit. Only half the field is available in HCON3 and that doesn’t include the region involved. (IRAS ran out of coolant before HCON3 could be completed).

    You can get a TAR file including the HCONs and the summed image at this link.

  4. Karl says:

    Thanks for your comments. Tom.

    We have been trying to locate technical information on IRAS Scan data (HCONS?) which appears to be the source data used By IRIS and SFD.

    Are there dates and observational position data available for these HCONS.

    I will follow your link and try to educate myself on retrieving these image and data files.

  5. Karl says:


    I retrieved the HCON files from IPAC and a astronomical FITS Viewer called FV. I found my (2) INFRARED Oblects. The images are horrible to view and the date information in the header said these images were taken in 1992. SkyView has much better image retrieval and we are all aware the images were taken in 1983 before IRAS Wan out of Nitrogen or whatever it was. Is there some other source for Time and Scan sequence information for the 1983 IRAS Mission.

    Any help would be appreciated

  6. Tom McGlynn says:

    Hi Karl,

    I’m posting from home and I don’t have the files in front of me. I just wanted to check that we are seeing the same things… The H0 image is the summed image (over all three HCON’s) and is what is used in SkyView surveys. It shows two bright objects.

    The H1 image is the earlier image and the H2 is supposed to be from a few weeks later. In both of these there is a single bright object which I believe is a planet moving from one position to the other. The H3 image is only about half filled and doesn’t include the region of interest.

    While I’m sure the information exists somewhere to recover the exact times of observations with IRAS it may take a while to uncover. The experts should be at LAMBDA and IRSA.

  7. Tom McGlynn says:

    Hi Karl,

    I finally found some Web sites which conveniently gave the position of the planets as a function of time (sad to say the easiest to use was an astrology site). Looking at these, it seems very likely that the two images are of Saturn taken in June/July 1983 in the HCON1 and HCON2. I haven’t tried to track things down exactly but Saturn’s seems to go right through where the two images are seen at about this time.


  8. Karl says:

    Hi Tom,

    I just recieved your E-mail.

    Thanks for your replies and all your effort.

    I have reviewed the HCON B2H0 B2H1 and B2H3 files. We are looking at the same images. I still have not confirmed any date or time data in the HCON files other than what looks like maybe a data release date from the early 1990’s. I have located the Object/s and it does appear to be Saturn as shown in Starry Night Pro around July 14, 1983.

    I am validating coordinates and dates now that you have taught me how to better use the HCON files and FV FITS Viewer.

    At this site, [link to pdsproto.jpl.nasa.gov] I also located the IRASPOS.TAB- This IRAS Position File describes the position of the satellite in its orbit about the Earth at 200-second intervals in heliocentric ecliptic rectangular coordinates (B1950) and SCAN.TAB- This IRAS Scan History File describes the pointing geometry and length of each survey-mode scan (by SOP OBS) over the entire mission.

    I am trying to deciepher them for use with Starry Night to get a simulated view of these coordinates on these dates.

    Is there a valid reason for NASA to allow what appears to be modification of existing Original IRAS Images in the IRIS and SFD Images?

    I found (6) image plates with different modifactions to the SkyView IRAS Images as found in the IRIS and SFD Images at this coordinate all at Skyview. There are still additional modifications, not shown in the SkyView images, on 3 different popular Internet based astronomy applications at this coordinate.

    Is this a common occurance with Astronomical Imagery provided to the public?

    What is the best source of unadulterated All-Sky INFRARED Images and how can that be validated?

    Thank you again for your time, effort and comments


  9. Tom McGlynn says:

    Hi Karl,

    Just to follow up on the issue of ‘NASA allowing’ this modification. The IRAS data are freely available to the public who may do anything they want with them. They are in the public domain. So anyone can modify them as they wish. E.g., the team that produced the IRIS data thought that they could do a better job of reducing some of the instrumental background and reprocessed the entire IRAS data set. For their purposes getting rid of a very bright solar system object seemed apropos and they were perfectly at liberty to do so. They made a substantial improvement on the original analysis. [This isn’t a knock on the original team. The IRIS analysis took computing resources that were prohibitively expensive at the time of the original IRAS analysis.] This kind of enhancement to the baseline data is one important goal of making the data sets available.

    A key issue here is the extent to which the authors describe the process they undertook. In the paper describing the IRIS survey (e.g., at
    ) the authors note that “We also dealt manually with Saturn because of its large extent in the affected HCONs”. So what they did was openly described. But it’s a lot easier to find this knowing what to look for than starting with an anomaly and trying to find the reason for it. Unfortunately, so far as I can see, they did not specify in the paper which HCONs this included. So it was hard for someone like you, starting with an HCON that seemed to be specially treated, to understand what had happened.

    Once the IRIS survey was published, its superior quality made it a natural candidate for inclusion in SkyView and with the team’s permission we added it.

    With regard to the ‘adulteration’ of astronomical images…. Almost all astronomical images — and certainly all images based on IRAS — involve a significant degree of software processing between the detector and the rendering of the image in a survey. IRAS was a scanning detector which essentially had photometers that measured the brightness of the sky in a little spot that moved along a narrow stripe (which in turn moved as the Earth revolved around the Sun). Converting these measurements to images was no small task. In astronomy and other sciences raw data is processed for both analysis and presentation. A properly calibrated dataset with instrument artifacts removed in a standard format is far more valuable for research (and usually more elegant for display), than raw data in arbitrary units with strong instrumental signatures in a complex and often proprietary format. It is important that the process for transforming the raw data to calibrated results is both understood and repeatable.

    In SkyView, the 2MASS data are less highly processed, I believe. The images in SkyView may be pretty close to the raw CCD images. However due to the optical characteristics of the telescope, 2MASS images have a fair number of artifacts, (e.g., see http://www.ipac.caltech.edu/2mass/releases/first/doc/sec4_7.html), so they need to be looked at with care. 2MASS is in the near infrared and has much higher resolution than IRAS/IRAS (1″ pixels versus 90″). So perhaps that would be better for your purposes.

  10. Bill Bartmann says:

    Great site…keep up the good work.

  11. alex esteban says:

    To Tom,

    What is wrong with astrology sites? It’s all about numbers. The equation of life. Number 9.

    alex esteban

  12. Styela says:

    The epoch of these images could be compared with the position of Jupiter to see if the objects show up in the expected images.

  13. Tom says:

    Hi Styela,

    You’re probably referring to my original speculation that the anomalous images were of a solar system object “likely Jupiter”. If you read further you’ll note that it turned out that my guess wasn’t quite right. They were of Saturn which was at exactly these locations at the epochs of the observations.

Leave a Reply

Your email address will not be published. Required fields are marked *