If you are new to Computer Science and Computer Vision, I would suggest to find some MOOC courses on one of  hese topics:

hope it helps

but i want to build an andriod app..

The links above are generic, so they apply to any platform!

What you ask for is very hard. Companies like Yahoo! paid millions of dollars to buy startups working on the field, to acquire their know-how...

At least if you want to separate photos of houses from photo of Dr. House (of persons in general...) and similar tasks needing some "intelligence", heuristics, neural networks (with lot of photos pre-sorted for training) and so on.

At our tiny level, we can, for example, dispatch photo by finding the color mostly used (reddish photos from blueish photos, etc.) but doing more is very, very hard.
Even what some cameras do now, detecting visages in the photo to be taken (to adjust focus, I suppose), isn't obvious, although there are some Processing libraries allowing to do that.

As PhiLho explained Computer Vision is  very sophisticated topic which requires years of studies and trial and error. With Processing.org you can accomplish simple things. 

As the first step you may want to look at Shiffmans examples on motion detection, and color tracking, this should be a good strarting point to understand simple ways of "calibrating" images.

thank you very much dimkir , it's help me so much..
i'm sorry because maybe my explanation it's not clear enough, my english is bad hehehhee

You need to post a picture of what exactly you're trying to accomplish. It is not very clear what "leaf color chart" is and what you're gonna do with it. 

leaf color chart..

I propose using getPixels and setPixels so that you can manipulate the array directly. I haven't run this, but it should give you the idea:

 

int[] pixels = new int[width * height];

bmOut.getPixels(pixels, 0, width, 0, 0, width, height);

// scan through all pixels

for(int i = 0; i < width * height; i++) {

    // get pixel color

    // pixel = src.getPixel(x, y);

    pixel = pixels[i];

    A = Color.alpha(pixel);

    // apply filter contrast for every channel R, G, B

    R = Color.red(pixel);

    R = (int)(((((R / 255.0) - 0.5) * contrast) + 0.5) * 255.0);

    if(R < 0) { R = 0; }

    else if(R > 255) { R = 255; }

 

    G = Color.green(pixel);

    G = (int)(((((G / 255.0) - 0.5) * contrast) + 0.5) * 255.0);

    if(G < 0) { G = 0; }

    else if(G > 255) { G = 255; }

 

    B = Color.blue(pixel);

    B = (int)(((((B / 255.0) - 0.5) * contrast) + 0.5) * 255.0);

    if(B < 0) { B = 0; }

    else if(B > 255) { B = 255; }

 

    // set new pixel color to output bitmap

    pixels[i] = Color.argb(A, R, G, B);

}

bmOut.setPixels(pixels, 0, width, 0, 0, width, height);

www.seocanny.com