Looping through pixels in an image

[Bod MacNeil]

Hi, I'm trying to complete a challenge on HackThisSite.  The challenge is - there is a hidden message in the pixels in an image.  I need to loop through the pixels to find this message.  I've only started the challenge and I've ran into a problem.  Here is my code so far:

import javax.imageio.ImageIO;
import java.awt.Image;
import java.io.File;
import java.io.IOException;


public class RBG {
    File imageSource = new File("Prog2.png");
    public void rbg(){
        
        try{
        Image prog2 = ImageIO.read(imageSource);
        }catch(IOException ioe){
            System.out.printf("Error!: %s", ioe);
        }
        for(int y = 0; y < prog2.getHeight(); y++){
            
        }
    }
    
}

The problem is this line

 for(int y = 0; y < prog2.getHeight(); y++){ 

It doesn't recognise prog2, which is the image.  Am i using this wrong?    thanks

[Piet Souris]

hi  Bod,

what about the scope of a variable?    

[Bod MacNeil]

aaahhhh so my for loop needs to be inside the try block  thank you!  Its giving me a new error now but I'll try figure this one out before asking.  

[Piet Souris]

Or maybe simpler: put your Image outside the try block.

If I may give an advice, out of my own experience :

if you have a pixel(x, y), then x is the horizontal distance, i.e. the column, and y is the vertical distance, or the row. So, if you start looping from 0 to image.height, you are talking about the rows. Therefore, it is wise to name the loop variable 'row', instead of y. That has saved me from many index out of bound errors, where I messed up the columns and the rows, and thus the width and the height!

[Piet Souris]

Between Bods opening post and his first reply there was a reply about the scope of the Image variable, declared within the try-block. That reply got somehow lost in the transfer of this topc to another forum.

[Bod MacNeil]

Looks like your reply has been returned.  Thanks for the advice as I was probably have confused them also!  

To loop through the imagine pixels, the only thing I could find was using Raster() and BufferedImage().  Would this be the right route to go down or does anyone know of a method which only reads pixels?  I can't seem to find anything else.  I continue to write my code using these methods as I'm sure it can work this way.  

[Piet Souris]

A BufferedIamge is easy to use. When you read your image with ImageIO.read, you automatically get a BufferedImage. It has the method 'getRGB' that you can loop through. The int that you get from this method is in the form ARGB, with the alpha component in the bits 24-31. You xcan use the 'setRGB' to write to the BI.

[Bod MacNeil]

I've been reading into BufferedImage and RBG and have a question which might seem stupid.  I see that we only use 3 colours - blue, green and red in an integer value.  If all 3 are 255, does that mean the pixel is white?  As I'm trying to find white pixels in a picture.  

[Carey Brown]

Bod MacNeil wrote:I've been reading into BufferedImage and RBG and have a question which might seem stupid.  I see that we only use 3 colours - blue, green and red in an integer value.  If all 3 are 255, does that mean the pixel is white?  As I'm trying to find white pixels in a picture.  

Yes. Black is 0x000000 and white is 0xFFFFFF and red is 0xFF0000 and green is 0x00FF00 and blue is 0x0000FF.

[Piet Souris]

Agree, but don't forget that getRGB returns a 32-bit integer, with the bits 24-31 containing the alpha component. If that alpha is 255, then the pixel is fully opaque, if alpha is 0, then the pixel is completely transparant. Does a white pixel count as white in that case?

A simple check would be to first have a look at your BufferedImage (either in Java, or in Paint or similar), to see if it is easy to detect the hidden message. Success!

[Bod MacNeil]

I wish it was that simple :P  The hidden message is the space between all the pixels in the white dots in the picture.  So I need to loop through the picture, once It finds a white dot it stores the dots pixel location and moves on to the next.  I just have no idea how to get the program to detect a certain colour!

public void rbg(){
        
        try{
        prog2 = ImageIO.read(imageSource);
        prog2Pixels = new BufferedImage(prog2.getWidth(), prog2.getHeight(), BufferedImage.TYPE_INT_ARGB);
        
        for(int row = 0; row < prog2.getHeight(); row++){
            
            for(int x = 0; x < prog2.getWidth(); x++){
                int colour = prog2.getRGB(row, x);
              
            }
        }//end of for loops
           
        }catch(IOException ioe){
            System.out.printf("Error!: %s", ioe);
        }//end of try catch
      
    }// end of rbg method

[Piet Souris]

for(int row = 0; row < prog2.getHeight(); row++){
             
            for(int x = 0; x < prog2.getWidth(); x++){
                int colour = prog2.getRGB(row, x);
               
            }
        }//end of for loops

Remember what I said about not mixing the row and the column?  

Well, if you have obtaibed the colour, as you do above, then you could do:

int red = (colour >> 16) & 0xFF;
int green = (colour >> 8) & 0xFF;

et cetera, and then test for:

if (red == 255 && green == 255 && blue = 255) {
   prog2Pixels.setRGB(colour);
}

Or should you test for colour NOT being white?

[Bod MacNeil]

I actually had this bit of code

if (red == 255 && green == 255 && blue = 255)

before but deleted it as my program didn't work, so I figured that was the problem.  It was clearly something else in the code that was the problem.

I only need to check for white pixels.  This is the image here:


So once we locate the first white pixel at the top, I save the location to an int, and then carry on looping to the next white pixel.  I will have a mess around with what you've suggested.  Thanks

[Bod MacNeil]

I did a quick test.  My loops seem to be looping through every pixel as the number of pixels were right, but its not picking up any colours when I use this:

for(int column = 0; column < prog2Pixels.getHeight(); column++){
            
            for(int row = 0; row < prog2Pixels.getWidth(); row++){
                
                int colour = prog2Pixels.getRGB(row, column);
                int alpha = (colour>>24) & 0xff;
                final int red = (colour & 0x00ff0000) >> 16;
                final int green = (colour & 0x0000ff00) >> 8;
                final int blue = colour & 0x000000ff;
                
                System.out.println(red + "" + green + "" + blue);
                
            }
            

This return 000.  Which at first I thought was black?  But if its picking the black up then it must also be picking up the white, which it is not.  IS there something Im missing in this?  thanks

[Bod MacNeil]

I fixed the problem.  It was code I had written earlier and forgot to remove that was causing the problem!!

[Carey Brown]

Piet Souris wrote:Or maybe simpler: put your Image outside the try block.

If I may give an advice, out of my own experience :

if you have a pixel(x, y), then x is the horizontal distance, i.e. the column, and y is the vertical distance, or the row. So, if you start looping from 0 to image.height, you are talking about the rows. Therefore, it is wise to name the loop variable 'row', instead of y. That has saved me from many index out of bound errors, where I messed up the columns and the rows, and thus the width and the height!

Piet, I'll have to disagree with you on this. Having done quite a bit of graphics programming X and Y are used almost exclusively, and ROW and COLUMN are nowhere to be found. Event the getRGB() method takes X and Y. I don't quite understand why you were running into out-of-bounds issues when using X and Y.

[Piet Souris]

hi Carey,

that's okay. I made this remark based on my own experience, but of course someone elses experience may be completely different.

If you compare the getRGB(x, y) to, say, array[x][y], where array is a 2D array (yeah I know, array of arrays   ) then there is a difference.
In array[x][y] the x is the row, going 'from top to bottom', and y is the column, 'going from left to right'. When dealing with the 'gerRGB(x,y), the situation is exactly reversed. The x is going from left to right (i.e. it is the column, going from 0 ti the widrh - 1), and y is going from top to bottom, ie is about the height. going from 0 to height - 1.

On more than one occasion, dealing with non square images, I mixed up the x and y, letting the x go from top to bottom like in the array, and got my index out of bounds errors. Since then I use consistently the notation array[row][column] for 2d arrays, and getRGB(column, row) for images.

I can imagine people finding this very personal notation of mine confusing. But don't say I didn't warn!    

[Bod MacNeil]

I have managed to loop through the image and find the white pixels.  This is the method which works.  

 public void rbg(){
        
        try{
        prog2 = ImageIO.read(imageSource);
        
        
        }catch(IOException ioe){
            System.out.printf("Error!: %s", ioe);
        }//end of try catch
        
        //loop through each pixel in prog2Pixels
        for(int row = 0; row < prog2.getHeight(); row++){
            
            for(int column = 0; column < prog2.getWidth(); column++){
                
                int colour = prog2.getRGB(column, row);
                int alpha = (colour>>24) & 0xff;
                int red = (colour & 0x00ff0000) >> 16;
                int green = (colour & 0x0000ff00) >> 8;
                int blue = colour & 0x000000ff;
              
                
                if(red == 255 && green == 255 && blue == 255){
                    
                     rbgColours.add(counter);
                    
                }
                counter++;
            }
           
        }//end of for loops
        sortArray();
    }//end of rbg method

Whats happening here is once its finds a white pixel it saves the number in an ArrayList called rbgColours.  

The next thing I nee to do is convert these numbers to characters by using the method below.  We do this by measuring the distance between pixels and subtracting the first pixel location with the seconds.  

private void sortArray(){
        
        for(int i = 0; i < rbgColours.size()-1; i++){
            
            answer = rbgColours.get(rbgColoursCount +1) - rbgColours.get(rbgColoursCount); //get ASCII from pixels
            if(answer != 32){
            ascii = (char) answer; //convert ASCII value to char
            str[strArrayCount] += ascii;
            }
            if(answer == 32){
                strArrayCount++;
            }
            
            
        }
      
    }//end of sortArray method

Once I ran the program I discovered that the hidden message was in morse code.  So I made another method to convert the more code to characters.  This method is below and uses a HashMap.

private void sortArray(){
        
        for(int i = 0; i < rbgColours.size()-1; i++){
            
            answer = rbgColours.get(rbgColoursCount +1) - rbgColours.get(rbgColoursCount); //get ASCII from pixels
            if(answer != 32){
            ascii = (char) answer; //convert ASCII value to char
            str[strArrayCount] += ascii;
            }
            if(answer == 32){
                strArrayCount++;
            }
            
            
        }
         System.out.print(str[0]);
    }//end of sortArray method

The last method also works.  The problem I am having is with the middle method sortArray();  When I run the program I get Null before the morse code.  Any ideas why, or a way around this?  Thanks.

[Bod MacNeil]

I posted the swrong method last.  Here is the HashMap method used to convert the code to ASCII.

private String morseCode(String choice){
        
        
        morse.put(".-", "A");
        morse.put("-...", "B");
        morse.put("-.-.", "C");
        morse.put("-..", "D");
        morse.put(".", "E");
        morse.put("..-.", "F");
        morse.put("--.", "G");
        morse.put("....", "H");
        morse.put("..", "I");
        morse.put(".---", "J");
        morse.put("-.-", "K");
        morse.put(".-..", "L");
        morse.put("--", "M");
        morse.put("-.", "N");
        morse.put("---", "O");
        morse.put(".--.", "P");
        morse.put("--.-", "Q");
        morse.put(".-.", "R");
        morse.put("...", "S");
        morse.put("-", "T");
        morse.put("..-", "U");
        morse.put("...-.", "V");
        morse.put(".--", "W");
        morse.put("-..-", "X");
        morse.put("-.--", "Y");
        morse.put("--..", "Z");
        morse.put(".----", "1");
        morse.put("..---", "2");
        morse.put("...--", "3");
        morse.put("....-", "4");
        morse.put(".....", "5");
        morse.put("-....", "6");
        morse.put("--...", "7");
        morse.put("---..", "8");
        morse.put("----.", "9");
        morse.put("-----", "0");
     
       String morsePick = morse.get(choice);
        return morsePick;
    }//end of morseCode method

[Piet Souris]

Hi Bod,

I'm afraid you posted your last post in the wrong topic. But can you explain the meaning of the counter values of IIRC, 32, 45 and 46? Are these white pixels always situated at these distances, even when split over more rows? But we'll done, it looks fine. And in case the message is about the location of a treasure: just send me a pm!    

[Carey Brown]

Your calls to morse.put() should be put in a static initializer block. You don't want to be doing those puts over and over for each choice.

[Bod MacNeil]

Carey Brown wrote:Your calls to morse.put() should be put in a static initializer block. You don't want to be doing those puts over and over for each choice.

AH that makes sense! thanks

Piet Souris wrote:Hi Bod,

I'm afraid you posted your last post in the wrong topic. But can you explain the meaning of the counter values of IIRC, 32, 45 and 46? Are these white pixels always situated at these distances, even when split over more rows? But we'll done, it looks fine. And in case the message is about the location of a treasure: just send me a pm!    

haha no treasure unfortunately  Here is a screenshot of the challenge as I'm not very good at explaining it.  It makes sense of the shite dots and what I'm trying to do.  


I've completed it now and are now onto the next challenge.  So no doubt I'll be posting a lot about that one :P