Meteorite camera project Meteotux_PI

[OlvinJanoisin]

Project to capture long exposure night time sky events with camera module.

Link: https://sites.google.com/site/meteotuxpi/

[Ravenous]

That's a good meteorite picture in your gallery page!

(I also like the banner picture at the top - the raspberry that fell to earth)

[caracol_uk]

Hi Olvin,

This sounds like a very interesting program and could be very useful. I've set up my Raspberry Pi as an all sky camera and I'm using Raspistill to obtain images with the Pi Noir camera. However, I'm limited to 6s exposures by the camera firmware. The details for your program suggest it can take exposures between 10s and 3600s. I'm curious how this is possible if the camera firmware has a 6s limit. Does your program take short exposures and combine them together to produce the requested exposure length?

On a separate note, when I run meteotux -d 10 the JPG images created are all black. Do you have any idea why this may be happening? I'd have expected the image to be completely white (i.e. very overexposed) since it was taken in daylight. I'm running Raspbian kernel version 3.12.26+ with the latest firmware.

Regards
Simon

[OlvinJanoisin]

Hi Simon,

Software takes short 0.5s exposured images and stacks them. Extending camera internal exposure time do not bring any major advantage (At least on my tests?).
The camera module works bit odd (imho), I realized the same; images are totally black when they are really overexposured.
Please try camera module on some dark to see if you get any valid image out?

I'll add some switch to control internal exposure time to see the changes.

BR,OlvinJanoisin

[caracol_uk]

Hi Olvin,

Thank you for your reply. I'll carry out some tests this evening once it gets dark to see how it looks.

I have been collecting 6s exposures, but these only show the brightest stars. If I stack 100x 6s frames together it still doesn't show fainter stars because the 6s exposures are too short to register them at all. A stack of 50x 12s frames would have more noise than 100x 6s, but fainter stars would be detected. Would it be easy to add an option to your program to allow setting individual exposure length and the ISO level used? If so that would be fantastic. Ideally I'd like to use 6s exposures at ISO 800.

Regards
Simon

[caracol_uk]

Hi Olvin,

The good news is that the program works in darker conditions. However, a 2 minute stack of 0.5s images shows much less detail than a single 6s image.

Regards
Simon

[DaveH]

Interesting project, thanks for posting this.

I'm currently working on a wildlife trail camera for use at night. While that will have some IR LEDs and a PIR, I have been thinking that what I will end up with is a Pi NoIR camera in a weather proof box and a tripod mount. What I actually use it for will just depend on what software I chose to run.

With your software, I will try pointing it at the sky some nights, rather than hedgehogs.

[OlvinJanoisin]

I added possibility to change the exposure time from 0.5s up to 6s. This is just a test version.
Available in here https://sites.google.com/site/meteotuxpi/download/

First images on long exposure will be a bit dark/odd colored because camera needs time to settle for final brightness so please be patient.

parameters -e 6000 -d 120 should give nice pictures on dark after 2-3 images.

There was problem that exposure time could not be over 1s at least on kernel version "3.12.22+ #691", so please update your system.

Running program with -v flag will show the frame length and that should be about the same as given exposure time.

Many thanks for Simon to test the exposure feature!!

[OlvinJanoisin]

Finally small hole on clouds... I'm impressed by the camera module capabilities.
[Edit] Camera module is the normal camera module, not NoIR (which is more sensitive)

6 sec. exposure, duration 300 sec.

[yv1hx]

Finally small hole on clouds... I'm impressed by the camera module capabilities.
[Edit] Camera module is the normal camera module, not NoIR (which is more sensitive)

6 sec. exposure, duration 300 sec.

Hello OlvinJanoisin,

Great picture, thanks for sharing!

I'm planning to use my Rpi + Camera setup for catching lighting flashes, somebody here: http://www.raspberrypi.org/forums/viewt ... 04#p634204 has recommended the use of Meteotux_PI, but I was reluctant to use it because I have the normal camera module (not the PiNoir version), but your results have encouraged me to have Meteotux_PI in the "Software To Use" list.

[OlvinJanoisin]

Here is one nice tail of meteorite caught with normal camera module:

[DougieLawson]

That image link doesn't work on the forum. Try sticking your image on http://imgur.com

[tchiwam]

Hello,

I need to do something similar here....

I have 2 identical setups, one with NoIR and one with IR filter
Both have 4mm F2.0

My todo list:
-Raw to float
-Dead pixel mapping
-Dead pixels interpolation
-Median window filter
-Raw linearization
-Pixel leakage (sun / over exposed moon)
-Mechanical shutter (GPIO)
-Automated dark frame substraction
-Automated hot pixel mapping
-Tracking polarizer (GPIO)
-Sun mask
- >650nm channel IR + R - R
-Sensitivity mapping
-HDR stacking
- Mid BG bandpass filter..
- Lens ? widest possible F1.4 ? or fastest possible 185deg ? faster than 1.4 ?
- Lens coordinate mapping
- etc etc... Maybe all, maybe none ... Maybe I should just buy a proper usb 1/2" CCD, but then I'd miss half the fun...

Star trails with touch of northern lights

1sttry.jpg (34.42 KiB) Viewed 8931 times

Thanks a lot for the fun pieces of this puzzle...

PS: the green thing on the side is northern lights...
PPS: the redish thing on the other side is light polution from my place...

[fsb888]

HI olvin
I am new to raps pi
I have downloaded program
how do I get it to run?

[waz1642]

I'm getting a segmentation fault error message.

Wayne

[Boomer120]

Olvin,
This is a great program.
Any chance you can modify it to either terminate at the end of a time period or have the program go through only one iteration of the settings? This would allow the user to have more tailored control over the program.

[IanS]

Nice work. I also have an interest in astrophotography, although have not had the chance to do any recently.
Some time ago I wrote my own (PC based) astrophotography software to handle aligning and stacking images. The alignment routines were based on https://www.researchgate.net/publicatio ... Algorithms, although these days I would see if OpenCV could do it for me. You need to be able to identify image rotation as well as translation.

Image stacking works to bring out faint stars because the signal strength increases linearly with the number of images, but the random noise increases with the square root of the number of images, so the signal-to-noise ratio also increases in proportion to the square root of the the number of images. It does means I rarely bothered going beyond about twenty 30s exposures, as to double the image quality I would need 80 images. However, if the objects within the images are not aligned to within 1 pixel, the stacking does not work to bring out the faint objects.

p.s. They are not meteorites unless they reach the ground. While still in the air they are meteors, and while still in space they are meteoroids.