zwostream2RMS

This project is based on the zwostream repositories from Eddy Daoud and Asgaut Eng.

I modified the source for streaming the zwo asi raw pictures to a v4l2loopback device. Than RMS can grab the pictures from this video device for meteor detection.

Stream to a v4l2loopback device and grab it with RMS Meteor Capture:

requirements

Linux Mint (or similar), a mini pc. we have a Zotac Nano with 4 cores.

The ZWO ASI SDK is installed in the directory above, so the libs and headers are in ../lib/x64/ and ../include/.

opencv4

v4l2loopback

sudo apt-get install v4l2loopback-dkms v4l2loopback-utils
sudo modprobe v4l2loopback devices=2

day 7, break through !!

autofocus, best for snap a single image every minute

# stream the ASI174MM (cam nr 0) to /dev/video0 (v4l2loopback device)
./zwostream2RMS -E -n0 | ffmpeg -f rawvideo -pixel_format gray8 -vcodec rawvideo -video_size 1936x1216 -i pipe:0 -vf format=yuv420p -f v4l2 /dev/video0
# and snap a picture from the stream
ffmpeg -f v4l2 -i /dev/video0 -frames:v 1 output.jpeg

# stream the ASI178MC (cam nr 1) to /dev/video1 (v4l2loopback device)
/zwostream2RMS -E -n1 | ffmpeg -f rawvideo -pixel_format bayer_rggb8 -vcodec rawvideo -video_size 3096x2080 -i pipe:0 -vf format=yuv420p -f v4l2 /dev/video1
# snap a picture from the stream:
ffmpeg -f v4l2 -i /dev/video1 -frames:v 1 output.jpeg

Problem: the autofocus modus of the ASI SDK raises timeout errors (error 11) while video capture. So you will loose pictures in the final stream!

fixed exposure for hunting meteors in the night

./zwostream2RMS -e100 -g200 | ffmpeg -f rawvideo -pixel_format gray8 -vcodec rawvideo -video_size 1936x1216 -i pipe:0 -vf format=yuv420p -f v4l2 /dev/video0

and then use RMS to capture the video stream from '0', see RMS config file.

-g200 is maybe not the best value, I'm testing it. -e100 should be 0.1sec, so you get 10 pictures per second. This is a theory.

ToDo

• test the binning, maybe it's good for the 178mc
• performance, ffmpeg is warm up the cpu, so find a way to reduce the cpu usage by ffmpeg; maybe streaming of the asi 178mc is not a good idea
• better text write to the snap image, font is to small
• maybe we can use opencv for snap and the image in the video capture routine, this can give us a better performance

This is the way!