More Image Stuff

[Tech B]

Been a while since I posted anything here, so I decided to post what I've currently been working on.

I use ImageChops to get a difference between two consecutive frames from my webcam. I would like to replace the pixel values > rgb 50, 50, 50 with 255, 0, 0.

I thought about using im.getdata() but it is too slow for "live feed".
I'm thinking the answer is in Image.point()

A more detailed discription on what I'm doing is found here.

My code so far. Its almost what I need.

from VideoCapture import Device
import pygame, Image, sys, ImageChops, time
from pygame.locals import *

cam = Device()

res = cam.getImage()
res = res.size

pygame.init()
screen = pygame.display.set_mode(res)
pygame.display.set_caption("Predator Vision: by Tech B")

while 1:
    for event in pygame.event.get():
        if event.type == pygame.QUIT: sys.exit()

    #grab two images
    im1 = cam.getImage()
    #you will need to wait a sec before taking another image
    #if not, the screen will remain black and buggy
    time.sleep(.2)
    im2 = cam.getImage()

    #im2-im1 per pixel
    diff = ImageChops.difference(im1,im2)
    
    #convert image to pygame type and then display
    shot = pygame.image.frombuffer(diff.tostring(), res, "RGB")
    screen.blit(shot,(0,0))
    pygame.display.flip()

[Beat_Slayer]

I think you may use the diff data with the use 'load' to alter the pixels like a array of pixels.

Nice project.

Cheers and Happy coding

[Tech B]

Using load is still slow. I know PIL has a method for doing it a lot quicker. The point() is what I need, but I'm having issues on figuring out how to pass it functions, and I don't know anything about look up tables.

Maybe there is a way to create a new image based off of the pixel values?

[Beat_Slayer]

Can't you separate it by channels (RGB) and treat the data on those channels.

[Tech B]

load() did the job. Frame rate is still kinda slow, but I'm at least seeing results.

...
diff = ImageChops.difference(im1,im2)

n = diff.load()
for x in range(0,res[0],2): #every other pixel for speed
    for y in range(0,res[1],2):
        if n[x,y] > (25,25,25):
            n[x,y] = (255,0,0)
        else:
            n[x,y] = (0,0,0)
...

[Beat_Slayer]

A hint that makes a huge difference in another project was doing like this.

...
diff = ImageChops.difference(im1,im2)

n = diff.load()

GRAY = (25, 25, 25)
RED = (255,0,0)
BLACK = (0,0,0)

for x in range(0, res[0], 2): #every other pixel for speed
    for y in range(0, res[1], 2):
        if n[x,y] > GRAY:
            n[x,y] = RED
        else:
            n[x,y] = BLACK
...

I know it doesn't look that better, and in my case and more variables, but it helped, you can try it.

Cheers and Happy coding