Grab Your Own Images From NOAA Weather Satellites

Can you believe that [hpux735] pulled this satellite weather image down from one of the National Oceanic and Atmospheric Administration’s weather satellites using home equipment? It turns out that they’ve got three weather satellites in low earth orbit that pass overhead a few times a day. If you’ve got some homebrew hardware and post processing chops you can grab your own images from these weather satellites.

The first step is data acquisition. [hpux735] used a software defined radio receiver that he built from a kit. This makes us think back to the software-radio project that [Jeri Ellsworth] built using an FPGA–could that be adapted for this purpose? But we digress. To record the incoming data a Mac program called DSP Radio was used. Once you do capture an audio sample, you’ll need something to turn it into an image. It just so happens there’s a program specifically for weather image decoding called WXtoImg, and another which runs under Linux called WXAPT. Throw in a little post processing, Robert’s your mother’s brother, and you’ve got the image seen above.

[Hpux735] mentioned that he’s working on a post about the antenna he built for the project and has future plans for an automated system where he’ll have a webpage that always shows the most current image. We’re looking forward hearing about that.

55 thoughts on “Grab Your Own Images From NOAA Weather Satellites

  1. Radio Electronics had a really nice WEFAX project back in the late 1970s. Way back when I was just first getting started in electronics, I built the analog front end of the project and fed in into the cassette port of my Apple ][plus.

  2. Decades ago, there was an article in Rainbow Magazine with code for a program for the Tandy Color Computer that would take shortwave input over the cassette adapter and turn it into one of these. Blew my fragile little mind, way back in something like 1982.

  3. I can believe it. I get my own weather images using WxToImg as well.

    The hardware involved does not need to be anything too complicated. An old police scanner will work. Basically, you just need to be able to tune into 137.100 MHz FM and have a half-way decent antenna for receiving. (Check out “The Zapper” and Quadrifilar Helix designs.

    Guess I need to start pumping more of my ham radio knowledge into HackADay – didn’t know you guys cared for such geekery. Also, checkout HamNation too.

  4. I remember a single opamp circuit that plugged into the parallel port that would decode weathersat, pocsag paging, and a few other modes. I probobly have one laying around somewhere, but if I remember right, it was a DOS only program and woldn’t work on anything after WIN 95.

  5. I used to dabble in this, but since the internet now has such images available there’s really little point, with the only issue being that each time you have a good link of live high resolution images they close it down after a month or so, it’s all paid by taxpayer money but oh how protective they are, but I guess the issue is as more and more people discover a link the bandwidth gets too much?

    1. No point?

      Might the pictures he receives be more up to date? Might it give him a greater variety of pics, see them before somebody decided which ones were important enough to post and which weren’t? And you did mention the url changing issues.

      More importantly though, since when did everything worth doing / building have to have a point? How many projects do we all ooh and ahh over that somebody could have filled the same need using the internet or some off the shelf device? That isn’t the point!

      1. It does have a certain excitement to get it directly, and indeed you might get more raw data, but I was explaining why for me the impetus had dissipated, the internet combined with the been-there-done-that effect.

        Long ago you could get weather images from shortwave btw, not just the satellite stuff but marine stuff with airpressure lines across areas, and they also had transmission of press photos on shortwave which you could intercept way back, and obviously that too is more exciting if you can’t get them on the internet.
        All that was phased out for the most part though, but there might be some die-hard amateurs sending shortwave ‘TV’ left (I wonder if that’s still monitored by oppressive regimes at all).

    2. “This can take a variety of forms, but the simplest is the QSD, or Quadrature Sampling Detector.”

      No. The simpleist is just an antenna straight into an A/D converter (usually just a soundcard mic input). It only gets half the bandwidth of QSD and without any amplifier I’m not sure about sensitivity but it IS simpler.

      Also… I wouldn’t recommend hooking an antenna directly to a soundcard that you can’t easily replace such as one built into the motherboard or an expensive one. Otherwise you at least want some sort of buffer circuit to protect the soundcard from static spikes on the antenna.

      Here’s an example of an SDR program that works this way: Unfortunately it’s an old project and you need to go to the wayback machine to find it. I’m sure there are others but that’s the one I know about.

    3. Another tip… I have read (but not tried this). Many soundcards have a capacitor in series with their input which cuts off the highest frequencies. For audio use this gets rid of hiss. For SDR use short that out and you get more bandwidth.

      1. A sound card is not going to sample a VHF signal, so connecting an antenna to your sound card gets you *nothing*.

        The link you mention is for a VLF (very low freq) receiver. VLF != VHF.

        What you might be able to do, however, is to mix a local oscillator signal with the incoming VHF signal to down-convert.

        If you phase-shift the local oscillator by 90 Deg, and feed the zero and 90 degree phase-shifted signal to two different mixers, you end up with outputs of I and Q. These could be fed to your sound card port for software decoding

      2. This is nonsense. A series pass capacitor is going to attenuate low frequencies, not high. The cap may be there to provide DC isolation, so if you decide, for some reason, to “short” it, you could create a host of other problems for yourself.

        If what you are referring to is a low pass filter, bear in mind that sound cards have to have a low pass filter someplace ahead of the A/D to prevent aliasing.

        I don’t see how disabling this gets you “more bandwidth.” Your bandwidth is set primarily by your sampling rate.

    1. Bandwidth? The bandwidth is not an issue because you’re only receiving data. It’s not like your antenna is going to soak up all that radiated energy from the satellite and not let anyone else have any. :)

      1. For POES and GOES analog signals, you need to adjust the discriminator on your receiver for a bit of extra bandwidth on the final stage. Otherwise, all you’ll get is a black screen on the decoder because all you are hearing is the carrier.

  6. Yeah, all it took was a small handheld aircraft radio with a bent coat hanger antenna in the 137-138 MHz band. The data was a 2400Hz carrier tone with the data AM modulated onto it. There were 2 point detectors, one visible, one IR. The rotation of the satellite was the line scan and the movement of the bird was the raster. A complete rotation would send a line of visible, a line of IR, and a line of image brightness cal and housekeeping data sent as scanned characters. And yes, the programming to turn the sat Keplerian orbital data and time into sat location so that political boundaries could be overlaid on the images was programming hell. Polar orbiters are needed because geosynchronous weather sats can’t see clouds at the poles.

  7. Nice thing to do on the weekend. I did this about 10 years ago and used a yagi fed into a TV rf amplifier fed into an rf scanner and finally the output into the computer sound card. I believe I used WXSAT for decoding. Never bothered adding colours though. Got some nice pictures of storm fronts heading into the British isles. Really nice to see amateurs still using these satellites :)

  8. Wow we are doing things that was normal in the 1980’s and calling it a hack.

    The building of the radio is noteworthy not the “ZOMG PICTURES FROM THE SKY! OMG!OMG!OMG!OMG”

    1. Don’t be so critical!

      I remember this too but how many people are there here that weren’t even alive in the 80s? I think posts like this are great! Show the kids you can make something without an Arduino!

  9. I was going to build one of these in the ’70s (without the computer!) but heard that they were going to drop the polar orbiters when the geosync satellites came online. I guess they changed their minds about that. :)

      1. I hope you mean funcube dongles are hard to get ahold of so you’re better off with the dedicated receiver. Because suggesting a $230 single use board being a better choice than a $180 60-1700mhz sdr is hilarious.

  10. Good job man! I like the different use of equipment and the image is pretty good.

    Did you happen to build a QHA antenna? Back in May of last year I did quite a bit of researching on how to build one. Lots of people had different ways but I found most write-ups to be lacking detail and quality. So when I built my own I made sure to share it (

    I eagerly await to see your antenna build!

    1. HAHAH!!

      Your site was one of the ones I used when I made mine! I didn’t copy it though, so it’ll still be a fresh take on it. I use #10 copper ground wire from the hardware store rather than the 3/8 soft copper tubing.

      1. Fresh take you say? Even better! :) Congratz on your successful build and thank you for sharing!

        I like the ground wire idea! The copper tubing on my antenna was difficult to work with and I don’t know if I got the bends quite right. I’ve seen some other people use coax through some small with good success as well.

  11. I’ve been doing this since the early 90’s — a radio and cheap computer were a lot easier to come by than a decent rural Internet connection in those days. It was tough being a meteorology student and living in the middle of nowhere, having to come up with live data for doing homework on.

    I remember having a DX440, simple dipole, and a simple converter to do HFFAX with. I picked up morse, FEC, and radiofax weather reports from all over the world. My favorite catch was from McMurdo, of all places, but my most regular datastreams were from WLO and Offut.

    These days, I’m no longer a meteorologist, so I do this just for fun. Got a laptop hooked up to a handheld scanner with the discriminator filter bypassed (gets wider bandwidth on the final) and one of Dr. Taggert’s “Zapper” antennas. I get 5 passes a day, on average, from NOAA-19. One of these days, I’ll get around to reprogramming the setup to scan all 4 NOAA POES birds and Meteor 3-2-3.

    My next project is going to be a GOES antenna and downconverter. Unfortunately, my construction skills are lacking, and I don’t feel like ponying up $$$ for the downconverter.

    Of course, if somebody wants a challenge, the normal stuff from POES and GOES are just low-res analog pictures. They also offer various datastreams, high-res digital pics, retransmissions of HFFAX data, etc.

  12. Actually this was first done in the 1960s (in an era of stone knives and bear skins). In the earliest hobbiest versions surplus World War II VHF radios were modified to receive the 137 Mhz. signals. An analog converter with a timing circuit changed the audio into X-Y coordinates and send them to a scope. A Polaroid camera was aimed at the scope and set for a long exposure image. As the satellite passed overhead it would plot dots on the scope, but it wasn’t until the pass was over and you developed the film (anybody remember B&W Polaroid film and the fixative sticks?) that you would see whether or not you had successfully gotten an image.

    Now it’s trivial – a VHF scanner with one of the crystals replaced with a capacitor to increase the bandwidth, a homebrew antenna made out of coat hangers, and a computer with a sound card.

    You should be aware that the new weather satellites, including the one scheduled to launch next week, do not have APT (Automatic Picture Transmission) anymore. There’s a new digital standard for the transmissions. I think there’s data available on the NOAA website. But the existing APT transmitters on the current satellites will continue to operate for many years before they’re completely replaced by the next generation satellites.

  13. I Just started doing this with a scanner for the past month, (Uniden BCT15XT) getting very good results.
    Most of the problem is poor signal but can be corrected in software (WxToImg is very good at this!)

    If anyone is interested:

    1) track the satellite on (click NOAA at the top)
    2) note the time, but click to show NOT visible passes. also note the Elevation; between 55 – 80+ are considered good passes for a decode!
    3) NOAA satellites use the following APT frequencies: (Narrow FM)

    noaa 19 = 137.1000 NFM
    noaa 18 = 137.9125 NFM
    noaa 15 = 137.6200 NFM
    Noaa 17 = 137.5000 NFM

    4) Get your self a copy of Audacity (Free), or use windows sound settings for the Line-in. I like using Audacity for watching the signal for quality.

    5) Also download APTDecoder (Free) watch the sound levels as this will affect the decoding quality, use the meter in the program to adjust this. Then keep on eye on Audacity for the levels, use the slider for the line-in and match this with the quality of the decode in APTDecoder.
    You can also import your WAVs into WxToImg for amazing results! (Saved for you from APTDecoder)

    1. Here’s the polar orbiting bird status —

      In summary:, only 3 APT birds left, no working LRPT units aloft, and only 3 HRPT units working. The new class of polar orbiters can’t fly soon enough, IMHO.

      POLAR APT/LRPT REPORT — 24th October 2011
      Satellite Frequency (MHz) Status Image Quality

      NOAA 15 137.620 (APT) on good. See Note 8.
      NOAA 17 137.500 (APT) on No images. See Notes 7&8.
      NOAA 18 137.9125 (APT) on good
      NOAA 19 137.100 (APT) on good but see Note 6
      Metop-A 137.100 (LRPT) off see Note 3
      Meteor M N1 137.100 (LRPT) sporadic see Note 6

      POLAR HRPT/AHRPT REPORT — 24th October 2011
      Satellite Frequency Antenna Mode Image Quality
      NOAA 15 1702.5 Omni HRPT Weak
      NOAA 16 1698.0 RHCP HRPT Good ( note 3)
      NOAA 17 1698.0 RHCP HRPT No images. See note 7.
      NOAA 18 1707.0 RHCP HRPT Good
      NOAA 19 1698.0 RHCP HRPT Good
      Feng Yun 1D 1700.4 RHCP CHRPT see note 8.
      Feng Yun 3A 1704.5 see note 5.
      Feng Yun 3B 1704.5 see note 5.
      Metop-A 1701.3 RHCP AHRPT Good (note 4)
      Meteor M N1 1700 See note 6.

  14. The Softrock receiver used for this project is almost IMPOSSIBLE to obtain because demand is so high and the project-master refuses to consider outside fulfillment for the kits. I consider the Softrock SDR project essentially dead as a result.

    1. Are you really going to call a project that ships 100-150 kits a week dead? It is hard to get one, and you have to check the website every so often. Big deal, don’t be so dramatic. The guy runs on EDT/EST since he lives in Indiana. You might have to adjust when you are checking because of that. It’s one guy kitting these and he still manages to do an estimate of 5,000-7,000 a year.

      Sign up to the Softrock40 list, sometimes when the radios are available someone will point that out on the list.

      Incidentally, as I am posting this the $20 Lite receiver and the $74 RXTX transceiver are both available. Right now.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.