cv::Mat/Numpy/Kinect over RTMP

asked 2014-01-13 15:01:31 -0500

MRDaniel gravatar image

Hello,

I have a linux box running rtmplite and i'd like to send cv::Mat/Numpy/Kinect data to my RTMP server.

(https://code.google.com/p/rtmplite/)

This aspect of internet communication seems to have a gap in it's tutorial base. I am able to use a .swf from the VideoIO (https://code.google.com/p/flash-videoio/) project to record and play videos from the server. How do we push Kinect data to the server?

1) Can we push a single image?

2) What method would be suitable for both depth and image frames?

3) Skeleton pionts offer another data source, although a json request may be suitable although they would need to combine together again at some stage on the server side.

Any other solutions available? Js-node? .NET? Rstp?

Regards,

Daniel

edit retag flag offensive close merge delete

Comments

librtmp-python

import librtmp import numpy

Create a connection

conn = librtmp.RTMP("rtmp://localhost:1935/myapp/file1", live=false)

Attempt to connect

conn.connect() stream = conn.create_stream()

Read 1024 bytes of data

data = stream.read(1024) print data

Create a connection

conn2 = librtmp.RTMP("rtmp://localhost:1935/myapp/file2.flv", live=True)

Attempt to connect

conn2.connect()

Get a file-like object to access to the stream

stream = conn2.create_stream()

Read 1024 bytes of data

stream.write(numpy.zeros((10,10),numpy.uint8))

The comments have been written as // instead of the usual python hastag, as it formats this comment incorrectly.

MRDaniel gravatar imageMRDaniel ( 2014-01-13 20:44:10 -0500 )edit