Tracking movement in certain area

General Discussion Other

extreal

Tracking movement in certain area

in General Discussion • Other • 1 year ago

Hi all,

I'm trying to make a program which tracks the amount of movement in a certain area through a webcam.Currently I've got the amount of movement occurring on the entire screen. Now I would like to track the movement in, let's say a rectangle of 200 by 200 pixels in the bottom right of the screen.

I've tried using a for loop, but since this loop just counts the pixels from left to right, top to bottom this would not give me the square area I'd like to track.

Thanks in advance for any help :)

Replies(3)

amnon.owed

Re: Tracking movement in certain area

1 year ago

Use get(). Do whatever operation you were doing to the selected PImage instead of the whole screen.

extreal

Re: Tracking movement in certain area

1 year ago

I have no idea what to do with get() in this code, the processing reference isn't being to clear about it either.

Currently I'm using this code to track the difference of each consecutive frame and add a score when there's enough movement:

for (int i = 0; i < cam.pixels.length; i++){
diffR = abs(red(cam.pixels[i]) - red(photo.pixels[i]));
diffG = abs(green(cam.pixels[i]) - green(photo.pixels[i]));
diffB = abs(blue(cam.pixels[i]) - blue(photo.pixels[i]));
diff.pixels[i] = color(diffR, diffG, diffB);
if (diffR > 105.0 && diffG > 105.0 && diffB > 105.0) {
if (timeleft > 0) {
score++;
}
}
}

amnon.owed

Re: Tracking movement in certain area

1 year ago

Something like this...

Pseudo-code Example

PImage currentImage, previousImage, diffImage;
void setup() {
currentImage = createImage(200, 200);
previousImage = createImage(200, 200);
diffImage = createImage(200, 200);
}
void draw() {
currentImage = cam.get(640-200, 480-200, 200, 200);
for (int i=0; i<currentImage.pixels.length; i++) {
diffR = abs(red(currentImage.pixels[i]) - red(previousImage.pixels[i]));
diffG = abs(green(currentImage.pixels[i]) - green(previousImage.pixels[i]));
diffB = abs(blue(currentImage.pixels[i]) - blue(previousImage.pixels[i]));
diffImage.pixels[i] = color(diffR, diffG, diffB);
if (diffR > 105.0 && diffG > 105.0 && diffB > 105.0) {
if (timeleft > 0) {
score++;
}
}
}
previousImage = currentImage.get();
}

p.s. If the score++ is in the for loop of the pixel array, I guess it quickly goes up when there is movement. Even for a 200x200 segment you already have 40000 possible +1's for every single frame.