working on a motion detection script for my camera

gistfile1.py

import os
import sys
import time
import Image
import base64
import urllib
import urllib2
import datetime
import cStringIO
import ImageChops
import numpy as np

# globals
DEFAULT_DEVICE_WIDTH  = 640
DEFAULT_DEVICE_HEIGHT = 480
SAVE_PATH = "/home/jbowman/cam/pics/"
THRESHOLD = 900
DEBUG = False

cam = {
        "user": "admin",
        "passwd": "",
        "url": "http://192.168.1.10/snapshot.cgi",
        }

# This class is based of a class posted on stackoverflow at the following
# url: 
# http://stackoverflow.com/questions/5524179/how-to-detect-motion-between-two-pil-images-wxpython-webcam-integration-exampl
class Images:
    def __init__(self, request, image1=None):
        self.image1 = image1
        self.request = request
        if not image1:
            self.image1 = self.getImage()
        time.sleep(0.5)
        self.image2 = self.getImage()

    def DoComparison(self, image1=None, image2=None):
        if not image1: image1 = self.image1
        if not image2: image2 = self.image2
        diffs = ImageChops.difference(image1, image2)
        return {
            "entropy": self.ImageEntropy(diffs),
            "img": image2
            }

    def ImageEntropy(self, image):
        w,h = image.size
        a = np.array(image.convert('RGB')).reshape((w*h,3))
        h,e = np.histogramdd(a, bins=(16,)*3, range=((0,256),)*3)
        prob = h/np.sum(h)
        return -np.sum(np.log2(prob[prob>0]))

    def getImage(self):
        imgRequest = cStringIO.StringIO(urllib2.urlopen(self.request).read())
        return Image.open(imgRequest)

if __name__=="__main__":
    base64string = base64.encodestring('%s:%s' % (cam["user"], cam["passwd"])).replace('\n', '')
    request = urllib2.Request(cam["url"])
    request.add_header("Authorization", "Basic %s" % base64string)
    imgCheck = {"img": None, "entropy": 0.0}
    while True:
        imgCheck = Images(request, imgCheck["img"]).DoComparison()
        if DEBUG:
            print imgCheck["entropy"]
        if imgCheck["entropy"] > THRESHOLD:
            now = datetime.datetime.now()
            savePath = "%s/%s/%s/%s" % (SAVE_PATH, now.year, now.month, now.day)
            if not os.path.exists(savePath):
                os.makedirs(savePath)
            imgCheck["img"].save("%s/%s.%s.%s.jpg" % (savePath, now.hour, now.minute, now.second), "JPEG")

I edited it to tweak the threshold to .07 from .08 after walking outside and it didn't see me. Seems like a person doesn't make much of a change at night, especially with my black jacket. I am getting false positives now. I think I may consider setting up a security light so I can increase the threshold.

New edit, changed the whole comparison. Now requires numpy. This is based off of the second example in the mentioned url. The variance in numbers for changes is greater so I think it's easier to detect stuff. I think I might have to start saving the entropy number as well for each image for a bit to get a better idea of what to tune for. Right now some shadows are kicking it off