You have to be more precise in your specification: the details might be clear in your head, but must be specified for others, as long as you are not alone to work on it.
For example, what is a cell? Have you an object associated to it?
What is a selection? Is that a simple boolean? Must it be an array parallel to your main array, or a field in the object associated to each cell?
What are the selection criteria? Must the cell be completely covered by the shape or only partially is enough? Is there a minimal coverage ratio needed?
Can the shape be totally arbitrary or can it have some predefined geometry?
What do you mean by " I have all coordinates of points from the origine (0,0)"?

Currently, it isn't really a programming question (I see them more as questions on an existing code, or more precisely specified), more like a math (geometry) question.

depending on the size of the array, you could make a mask, from a PGraphics object.
you draw the shape (or multiple shapes) on a PGrahics object (Offscreen) which has the same dimensions as your 2Darray has, ... with a different color than the background. after that, you iterate over all the pixels, and extract the ones, that have the color of the shape. the indices of these pixels would be the same as the indices of your array.
this way, you can make more complex shapes too!

but this only makes sense if your array is not too big, otherwise it gets really slow. in that case you need a smarter solution, which of course is more complex.

some hints:
bresenham, aliasing

First, thank you all, for your answers.

This is actually a little part of my thesis work (MSc. in Architecture).




@ Phil,

*"Cell" (in image, simple yellow square): Every square object that you see in the image. I have a 2D array of them  [i,j]
like:

for (i = 0 ; i < 8 ; i++) {
      for (j = 0 ; j < 8 ; i++) {
            rect(i*50,j*50,50,50);
   }
}



*selection zone (red zone in the image) is a quadrilateral wich is difined by 4 points (w1,w2,w3,w4). these points are calculated by an other algorithm wich calculates their x and y distance from the origine (0,0)  -top left in the image-.

*Every cell or let's say square have a color characteristic, for ones which are out of the selection zone will be filled black and the rest -wich are in the selection zone- will be filled white.

* Selection criteria: starting point of the square (i*50,j*50) must be in the zone.
In fact I would love to define as cell must be completely in the selection zone but I guess it will be too complex for me.


Phil, I hope things are more clear, the whole project is in 3D but for this part I try to simplify as possible as I can.




@Thomas,

* things you mentionned are so cool but my whole project have lots of complex equations and this part is just an implantation so I would like to keep it as simple as I can.

I need really the simplest solution.



thank you in advance for your answers.

Thank you Thomas, that is very helpful.

A question for my thesis; is this code based on Bresenham's line algorithm?

thank you again.

no, the example above just uses processings standard drawing technique.
(internally, im sure, its used somwhere)

some time ago i tried to make the bresenham-algorithm in processing:
http://thomasdiewald.com/processing/applets/applet_diewald_bresenham_algorithm_v4/

but i think this just makes it more complicated than it should be.