ffi-libfreenect¶ ↑

FFI-based Ruby wrapper for the OpenKinect libfreenect library.

Requirements¶ ↑

* ffi >= 0.5.0

Installation¶ ↑

via Gem¶ ↑

(sudo)? gem install ffi-libfreenect

via Rake¶ ↑

git clone http://github.com/jgrunzweig/ffi-libfreenect.git
cd ffi-libfreenect
(sudo)? gem install jeweler
rake install

Synopsis¶ ↑

require 'freenect'

ctx = Freenect.init()
devs = ctx.num_devices

STDERR.puts "Number of Kinects detected: #{devs}"
unless devs > 0
  STDERR.puts "Error: no kinect detected"
  exit 1
end

STDERR.puts "Selecting device 0"
dev = ctx.open_device(0)

dev.set_led(:red)   # play with the led
dev.set_tilt_degrees(30)  # tilt up to max
sleep 1
dev.set_tilt_degrees(-30) # tilt down to max
sleep 1
dev.set_tilt_degrees(0.0) # tilt back to center
sleep 1
dev.set_led(:green)   # play with the led

# Actual video and depth capture work similarly to eachother.
# But they're still both pretty un-sugary.
#
# Future versions will probably abstract this and try to make it more
# ruby-ish.
#
# The example below shows how to write a single video frame to a PPM file.

dev.depth_mode = Freenect.depth_mode(:medium, :depth_11bit)

dev.video_mode = Freenect.video_mode(:medium, :rgb)

dev.start_depth()
dev.start_video()

$snapshot_finished = nil

STDERR.puts "Taking snapshot"
dev.set_video_callback do |device, video, timestamp|
  if not $snapshot_finished
    fname = "%u.ppm" % timestamp
    STDERR.puts "Writing #{fname}"
    File.open(fname, "w") do |f|
      f.puts("P6 %d %d 255\n" % [ dev.video_mode.width, dev.video_mode.height ] )

f.write(video.read_string_length(dev.video_mode.bytes)) end

    $snapshot_finished = true
  end
end

until $snapshot_finished 
  break if (ctx.process_events < 0)
end

dev.set_led(:off)
dev.stop_depth
dev.stop_video
dev.close
ctx.close

Copyright¶ ↑

Copyright © 2010 Josh Grunzweig & Eric Monti. See LICENSE.txt for details.