Raspberry Pi NOAA and Meteor-M 2 Receiver

This instructable will help you set up a receiving station for not just APT from NOAA-15, 18 and 19, but also Meteor-M 2.

It's really just a small follow-on project to haslettj's great "Raspberry Pi NOAA Weather Satellite Receiver" project.

Step 1: Start With Haslettj's Instructable

Before you follow haslettj's Instructable, note these changes first:

The old address for wxtoimg is no longer up. You can now use this address instead in the wget command:

https://www.wxtoimgrestored.xyz/beta/wxtoimg-linux-armhf-2.11.2-beta.tar.gz

I also had to make a change to the "receive_and_process_satellite.sh" script, because rtl_fm apparently does not actually output "wav" format audio. But it's no problem, sox can handle it. So I changed this line:

sudo timeout $6 rtl_fm -f ${2}M -s 60k -g 45 -p 55 -E wav -E deemp -F 9 - | sox -t wav - $3.wav rate 11025

To this (but remember to replace the "-g 0" gain setting and "-p 68" PPM frequency error setting with something that works for your hardware):

sudo timeout $6 rtl_fm -f ${2}M -s 48k -g 0 -p 68 -E dc -A fast -F 9 - | sox -t raw -r 48000 -es -b16 -c1 -V1 - $3.wav rate 11025

In the same script, you may also want to change the wxtoimg argument "-e ZA" to "-e MSA" to get nice color images, like this:

/usr/local/bin/wxtoimg -m ${3}-map.png -e MSA $3.wav $3.png

Now go do the Instructable!

https://www.instructables.com/id/Raspberry-Pi-NOAA...

Step 2: Install GnuRadio and Blocks for RTL-SDR

The Meteor-M 2 receiver uses GnuRadio. To install what you need, do this:

sudo apt install gnuradio

sudo apt install gr-osmosdr

Step 3: Download GnuRadio Scripts

If you're not familiar, GnuRadio includes a graphical tool called GnuRadio-Companion which can be used to build flow graphs and compile them into Python code which is then executed.

I have forked "otti-soft"s "meteor-m2-lrpt" receiver by modifying some of the parameters to improve performance and using RTL-SDR instead of Airspy. Download it from here:

https://github.com/NateDN10/meteor-m2-lrpt

The .grc files can be opened with GnuRadio-Companion, but they're not executable scripts - they're there for your reference and to play around with. To just get it working, copy the file "rtlsdr_m2_lrpt_rx.py" into your /home/pi/weather/predict directory, and make sure it's executable:

chmod +x rtlsdr_m2_lrpt_rx.py

You'll also want to change the frequency offset:

self.rtlsdr_source_0.set_freq_corr(69, 0)

And gain to whatever works for your setup:

self.rtlsdr_source_0.set_gain(4, 0)

Step 4: Download the Decoder

Download "artlav"s Meteor LRPT decoder from here - you want the Linux ARM version:

http://orbides.org/page.php?id=1023

You can accomplish this on the Raspberry Pi using these commands:

cd /home/pi/weather

wget https://orbides.org/etc/medet/medet_190825_arm.tar.gz
mkdir medet; cd medet
tar xvzf ../medet_190825_arm.tar.gz

You should now have a directory called "medet" within your "weather" directory, and inside it should be a "medet_arm" executable.

Step 5: Download Other Tools

To fix the aspect ratio of the images we're going to use dbdexter's "meteor_rectify" Python tool from Github.

If you don't already have git and ImageMagick installed:

sudo apt install git
sudo apt install imagemagick

Then clone the repository:

cd /home/pi/weather

git clone  https://github.com/dbdexter-dev/meteor_rectify.gi...

You may also need the "pillow" and "numpy" Python libraries:

pip3 install numpy

pip3 install pillow

Step 6: Update the Scripts

First, add the following line to the end of "schedule_all.sh":

/home/pi/weather/predict/schedule_satellite.sh "METEOR-M 2" 137.1000

Then, in "schedule_satellite.sh", change this block:

if [ $MAXELEV -gt 19 ]; then
    echo ${1//" "}${OUTDATE} $MAXELEV

    echo "/home/pi/weather/predict/receive_and_process_satellite.sh \"${1}\" $2 /home/pi/weather/${1//" "}${OUTDATE} /home/pi/weather/predict/weather.tle $var1 $TIMER" | at `date --date="TZ=\"UTC\" $START_TIME" +"%H:%M %D"`
fi

To this:

if [ $MAXELEV -gt 19 ]; then
    echo ${1//" "}${OUTDATE} $MAXELEV

    if [ "$1" == "METEOR-M 2" ]
      then
        echo "/home/pi/weather/predict/receive_and_process_meteor.sh \"${1}\" $2 /home/pi/weather/${1//" "}${OUTDATE} /home/pi/weather/predict/weather.tle $var1 $TIMER" | at `date --date="TZ=\"UTC\" $START_TIME" +"%H:%M %D"`
      else
        echo "/home/pi/weather/predict/receive_and_process_satellite.sh \"${1}\" $2 /home/pi/weather/${1//" "}${OUTDATE} /home/pi/weather/predict/weather.tle $var1 $TIMER" | at `date --date="TZ=\"UTC\" $START_TIME" +"%H:%M %D"`
    fi
fi

Finally, create a new script called "receive_and_process_meteor.sh" with the following contents:

#! /bin/bash
# $1 = Satellite Name
# $2 = Frequency
# $3 = FileName base
# $4 = TLE File
# $5 = EPOC start time
# $6 = Time to capture

cd /home/pi/weather

timeout $6 predict/rtlsdr_m2_lrpt_rx.py $1 $2 $3

# Winter
#medet/medet_arm ${3}.s $3 -r 68 -g 65 -b 64 -na -S

# Summer
medet/medet_arm ${3}.s $3 -r 66 -g 65 -b 64 -na -S

rm ${3}.s

if [ -f "${3}_0.bmp" ]; then
	#rm ${3}.s

	dte=`date +%H`

	# Winter
	#convert ${3}_1.bmp ${3}_1.bmp ${3}_0.bmp -combine -set colorspace sRGB ${3}.bmp
        #convert ${3}_2.bmp ${3}_2.bmp ${3}_2.bmp -combine -set colorspace sRGB -negate ${3}_ir.bmp

	# Summer
	convert ${3}_2.bmp ${3}_1.bmp ${3}_0.bmp -combine -set colorspace sRGB ${3}.bmp

	meteor_rectify/rectify.py ${3}.bmp

	# Winter only
	#meteor_rectify/rectify.py ${3}_ir.bmp

	# Rotate evening images 180 degrees
	if [ $dte -lt 13 ]; then
		convert ${3}-rectified.png -normalize -quality 90 $3.jpg

                # Winter only
		#convert ${3}_ir-rectified.png -normalize -quality 90 ${3}_ir.jpg
	else
		convert ${3}-rectified.png -rotate 180 -normalize -quality 90 $3.jpg
                # Winter only
                #convert ${3}_ir-rectified.png -rotate 180 -normalize -quality 90 ${3}_ir.jpg
	fi

	rm $3.bmp
	rm ${3}_0.bmp
	rm ${3}_1.bmp
	rm ${3}_2.bmp
	rm ${3}-rectified.png

	# Winter only
	#rm ${3}_ir.bmp
	#rm ${3}_ir-rectified.png
fi

Make it executable:

chmod +x receive_and_process_meteor.sh

And that's it! Next time your existing cron job runs to schedule satellites, Meteor-M 2 will be scheduled also. The decoder will output a .bmp using APID 66 for red, 65 for green, and 64 for blue.

The standard output from the scripts, when they're run by the scheduler, is appended to /var/mail/pi. To read it, use this command:

less /var/mail/pi

And to erase old messages, do this:

> /var/mail/pi

Comments

[ocellaris]

Kudos for the great write-up. I'm close to having it working, but having some trouble with the receive_and_process_meteor.sh script. On line 10, the script successfully creates the .s file:

-rw-r--r-- 1 pi pi 139485184 Oct 25 10:06 METEOR-M220201025-095055.s

On line 13, the medet_arm command creates the following 3 files:

-rw-r--r-- 1 pi pi 2911286 Oct 25 10:10 METEOR-M220201025-095055_0.bmp
-rw-r--r-- 1 pi pi 2911286 Oct 25 10:10 METEOR-M220201025-095055_1.bmp
-rw-r--r-- 1 pi pi 2911286 Oct 25 10:10 METEOR-M220201025-095055_2.bmp

Line 20 is an if statement that checks for the existence of "${3}.bmp". In my case ${3} is "METEOR-M220201025-095055" so the if statement is checking for "METEOR-M220201025-095055.bmp" which does not exist. Therefore the argument to the if statement is false. Is this a script error? Should the if statement be checking for one of the files that was generated?

Next I tried manually converting the generated .bmp files using commands from the script, but receive the following errors:

pi@satellite:~/weather/data/tmp$ convert METEOR-M220201025-095055_1.bmp METEOR-M220201025-095055_1.bmp METEOR-M220201025-095055_0.bmp -combine -set colorspace sRGB METEOR-M220201025-095055.bmp
convert-im6.q16: length and filesize do not match `METEOR-M220201025-095055_1.bmp' @ warning/bmp.c/ReadBMPImage/839.
convert-im6.q16: length and filesize do not match `METEOR-M220201025-095055_0.bmp' @ warning/bmp.c/ReadBMPImage/839.

pi@satellite:~/weather/data/tmp$ convert METEOR-M220201025-095055_0.bmp METEOR-M220201025-095055_1.bmp METEOR-M220201025-095055_2.bmp -combine -set colorspace sRGB METEOR-M220201025-095055.bmp
convert-im6.q16: length and filesize do not match `METEOR-M220201025-095055_0.bmp' @ warning/bmp.c/ReadBMPImage/839.
convert-im6.q16: length and filesize do not match `METEOR-M220201025-095055_1.bmp' @ warning/bmp.c/ReadBMPImage/839.
convert-im6.q16: length and filesize do not match `METEOR-M220201025-095055_2.bmp' @ warning/bmp.c/ReadBMPImage/839.

Any comments or suggestions would be greatly appreciated.

[NateDN10]

Check out the latest update to the script. This issue should be fixed now.

[solaria137]

"@ warning/bmp.c/ReadBMPImage/839."

It's the BMP files that medet creates that cause the message, but it's just a warning, so it can be ignored. Did the BMP's look OK otherwise?

I've noticed that output files generated by medet also have more black lines than the same S-file decoded using LRPT_Decoder_v56 (Windows). Need a better decoder I guess.

[VU2IIA]

After finishing steps in this instructable, there is another nice program to add map overlay just like MeteorGIS for raspberry pi /Linux by Viktor Venczel, https://github.com/Digitelektro/MeteorDemod. This will require you to make some changes to "receive_and_process_meteor.sh" and use "MeteorDemd"
My Meteor M2 image received using Pi4 and this instructable with MeteorDemod.
full res- https://qsl.net/vu2iia/11-jan-2021.jpg

VU2IIA

[WessexWeather]

Hello, thanks for sharing this. Would you care to share how you managed to get MeteorDemd installed, and the modifications you made to receive_and_process_meteor.sh?

[Callsign_M0NDE]

Has anyone used the Pi4 for this project?

[mardex1234]

yes, working for me

[tecibob]

Hi,
WXTOIMG images seem to work fine, however when I install the meteor stuff wxtoimg seems to not be able to get any pictures at all.
When the meteor satellite goes over a new file is created however, no end image is created. It is a file with a .s extension.

I am using raspberry pi 4 2gb and a Nooelec RTL-SDR USB device.

Thanks all!

[mardex1234]

I have the same problem now. Could you solve same?

[mardex1234]

Good day,
anyone an idea why my meteor images are all grey and this of a guy not far from me are coloured. Am I doing wrong decoding ? and what could be the solution=
Thanks

[deadlybynite]

Inside receive_and_process_meteor.sh on line 20, the script writes "if [ -f "${3}.bmp" ]; then" but medet always outputs three bmp files - ${3}_0 ${3}_1 and ${3}_2. There never is a ${3}.bmp making the statement always false and the rest of the code such as convert and rectify.py never run. Is this a typo, or am I missing something?

[ke8aff]

On the Meteor Side of things, where would one be able to specify a RTL-SDR Device Index in order to use a separate rtl-sdr device to decode Meteor M 2 ?

[aaz.off]

I have the same question

[mattybaus60]

There is Meteor-M2 and MeteorM2-2
both appear to have LRPT 72kbit
how would I add both satellites to your script mods - sorry not my area of expertise
but had a stab at it using

if [ $MAXELEV -gt 19 ]; then
echo ${1//" "}${OUTDATE} $MAXELEV
if [ "$1" == "METEOR-M 2" ]
then
echo "/home/pi/weather/predict/receive_and_process_meteor.sh \"${1}\" $2 /home/pi/weather/${1//" "}${OUTDATE} /home/pi/weather/predict/weather.tle $var1 $TIMER" | at `date --date="TZ=\"UTC\" $START_TIME" +"%H:%M %D"`
elif [ "S1" == "METEOR-M2 2" ]
then
echo "/home/pi/weather/predict/receive_and_process_meteor.sh \"${1}\" $2 /home/pi/weather/${1//" "}${OUTDATE} /home/pi/weather/predict/weather.tle $var1 $TIMER" | at `date --date="TZ=\"UTC\" $START_TIME" +"%H:%M %D"`

else
echo "/home/pi/weather/predict/receive_and_process_satellite.sh \"${1}\" $2 /home/pi/weather/${1//" "}${OUTDATE} /home/pi/weather/predict/weather.tle $var1 $TIMER" | at `date --date="TZ=\"UTC\" $START_TIME" +"%H:%M %D"`
fi
fi

have to wait a few hours for a pass to test

[tecibob]

Meteor m2-2 had a collision with a micro meteorite or some space debris and the 137mhz LRPT broadcast has been disabled.

[NateDN10]

Meteor M2-2 is no longer transmitting LRPT. It was struck by a micrometeorite and its cooling system and batteries were damaged. It still transmits HRPT in the daylight, at L-band.

[solaria137]

Got it to work, but ran into problems using Raspberry Pi. "rtlsdr_m2_lrpt_rx.py" gives the "corrupted double-linked list" error, which I guess is a python2/3 compatibility problem?

Fortunately, I've got an X86 (Intel(R) Atom(TM) x5-Z8500 CPU @ 1.44GHz) running Lubuntu that works fine... although 'predict' and 'ncurses' aren't in the repository and had to be built from source.

Looks like the camera on the Meteor is tilted forward a few degrees? unlike NOAA, which points straight down? Kind of gives a 3D perspective.

Captured from Southern California, showing the central valley and Baja California. I think the jagged white line is the Colorado River around the Grand Canyon.

Hope to add an LNA/filter in the next few days

[WessexWeather]

Is it possible to overlay landmasses and countries with these scripts? It would be particularly useful with the IR images, like this...

[erroltheterrible]

Look at "wxmap", it installed with wxtoimg. You can feed the created map png to wxtoimg and it will overlay the map onto the decoded image... You can also tell wxmap how much detail you want in the map.

[WessexWeather]

Thanks, so is this automatically installed with wxtoimg? Could this be the answer NateDN10?