motion computation from video using pyglet in python

Question

Hi, I am writing a simple motion detection program but i want it to be cross platform so im using python and the pyglet library since it provides a simple way to load videos in different formats (specially wmv and mpeg). So far i have the code given below which loads the movie and plays it in a window. Now i need to: 1) grab frame at time t and t-1 2) do a subtraction to see which pixels are active for motion detection.

any ideas on how to grab frames and to skip over frames and is it possible to put the pixel values into a matrix in numpy or something directly from pyglet? or should look into using something other than pyglet?

thanks kuaywai

---

import pyglet  
import sys  
window = pyglet.window.Window(resizable=True)  
window.set_minimum_size(320,200)  
window.set_caption('Motion detect 1.0')  
video_intro = pyglet.resource.media('movie1.wmv')  
player = pyglet.media.Player()  
player.queue(video_intro)  

print 'calculating movie size...'  
if not player.source or not player.source.video_format:  
    sys.exit  

myWidth = player.source.video_format.width  
myHeight = player.source.video_format.height  
if player.source.video_format.sample_aspect > 1:  
    myWidth *= player.source.video_format.sample_aspect  
elif player.source.video_format.sample_aspect < 1:  
    myHeight /= player.source.video_format.sample_aspect  

print 'its size is %d,%d' % (myWidth,myHeight)  
player.play()  

@window.event  
def on_draw():  
        window.clear()  
        (w,h) = window.get_size()  
        player.get_texture().blit(0, h-myHeight,  
                              width=myWidth,  
                              height=myHeight)  

pyglet.app.run()  

Answer 1

To me it seems like you have to skip the play function and manually step through the video/animation, maybe using source.get_animation().frames which is a list of frames where each frame is a simple image. I'm guessing this isn't really going to be pracitical with large videos but that is generally not something you should be handling in Python anyway.