I recently started playing around with pixel based collision detection, and it is working as I expected. Currently I use the pixel based collision to determine whether or not the player has hit a dangerous object (like a spike or such). All other "tiles" in my world are still being handled by rectangle based collisions. I would like to change that though.

What I do currently for collision with solid tiles is that I check the intersection depth, one axis at a time, and adjust accordingly. This works fine so far. But now I would like to be able to do pixel based collision responses on my tiles, which means that I can no longer use the intersection depth of two rectangles, because this would clearly give some weird results.

I can't seem to find much about pixel based collision responses online. All I actually want to know is how I get the intersection depth between two images...

The code I use for pixel based collision detection is the following:

private bool PixelbasedCollision(Rectangle rect1, Color[] data1, Rectangle rect2, Color[] data2)
        {
            int top = Math.Max(rect1.Top, rect2.Top);
            int bottom = Math.Min(rect1.Bottom, rect2.Bottom);
            int left = Math.Max(rect1.Left, rect2.Left);
            int right = Math.Min(rect1.Right, rect2.Right);

            for (int y = top; y < bottom; y++)
            {
                for (int x = left; x < right; x++)
                {
                    Color color1 = data1[(x - rect1.Left) + (y - rect1.Top) * rect1.Width];
                    Color color2 = data2[(x - rect2.Left) + (y - rect2.Top) * rect2.Width];

                    if (color1.A != 0 && color2.A != 0)
                        return true;
                }
            }
            return false;
        }

        private Color[] GetTileColorData(int x, int y)
        {
            Color[] tmp = new Color[TILE_SIZE * TILE_SIZE];
            _tileSheet.GetData<Color>(0 * TILE_SIZE, new Rectangle(_world[x,y].Variant * TILE_SIZE, (int)(_world[x,y].Type - 1) * TILE_SIZE, TILE_SIZE, TILE_SIZE), tmp, 0, tmp.Length);
            return tmp;
        }

Basically the GetTileColorData method creates an array of the Color based on an image. These are then given as arguments to the PixelbasedCollision method.

I was hoping that it was somehow possible to get the exact collision depth based on what happens in the PixelbasedCollision method, but I'm kind of stuck.

My question in short is:

How can I get the intersection depth between two images in a pixel based collision? - And is the information I need somewhere inside my PixelbasedCollision method?

Thanks for reading! :-)

I was sitting and thinking about this code (I can't test it right now) and thought of a way to check the intersection depth.

I found a possible solution. Wouldn't it be possible to just do the following:

y - top = vertical Intersection Depth

x - left = horizontal Intersection Depth

Since top and left defines the top left point for the rectangle that is created due to the Rectangle based intersection, y and x simply shows how many pixels into that "sub rectangle" we are.

This seems right in my head. Anyone who agrees or disagrees?

Thanks for reading!