Pixel by Pixel (and my attempts at Edge Detection)

For this weeks Computational Cameras assignment, I attempted to do edge detection. I looked at trying to combine the example code for doing pixel by pixel in video set out by Dan O’Sullivan here in tracking the single pixel / tracking a difference in color. I then attempted to use what is on the processing.org website for edge detection, but ended up with some rather strange results.

(Video is pending):

Here is the code (both of these results came from messing around with the initial kernel, though in the second one I maintained the information for all 3 colors rather than converting to grayscale). Oh yeah, and before I forget to mention it, Eclipse is super annoying

import processing.core.PApplet;
import processing.video.Capture;

public class n extends PApplet {

Capture video;// regular processing library

int threshold = 250; //255 is white, 0 is black

int aveX, aveY; //this is what we are trying to find

float objectR = 0;

float objectG = 0;

float objectB = 0;

boolean debug = true;

 public void setup() {

 size(640, 480);
 println(Capture.list());
 video = new Capture(this,width,height);
 //video.settings();

}

public void draw(){

 if (video.available()){
 video.read();
 //enter into the classic nested for statements of computer vision
 for (int row = 1; row < video.height-1; row++) {
 for (int col = 1; col < video.width-1; col++) {
 //the pixels file into the room long line you use this simple formula to find what row and column the sit in 

 int offset = row * width + col;
 //pull out the same pixel from the current frame
 int thisColor = video.pixels[offset];

 int lastColor = video.pixels[offset-1];
 //pull out the individual colors for both pixels
 float r = red(thisColor);
 float g = green(thisColor);
 float b = blue(thisColor);

 lastColor = row*width + col -1;
 objectR = red(lastColor);
 objectG = green(lastColor);
 objectB = blue(lastColor);

 //in a color "space" you find the distance between color the same whay you would in a cartesian space, phythag or dist in processing
 float diff = dist(r, g, b, objectR, objectG, objectB);

 if (diff > threshold) {  //if it is close enough in size, add it to the average
 video.pixels[offset-1] = color(255); 

 }
 }
 }

 }

 image(video, 0, 0);
}

public void keyPressed() {

 //for adjusting things on the fly
 if (key == 's') {
 video.settings();
}
}
}