The following is a function to "flood fill" a region on the active plotting device. Once called, the user will be asked to click on the desired target region. The flood fill algorithm then searches neighbors in 4 directions of the target cell (down, left, up, right) and checks for similar colors to the target cell. If neighboring cells are of the same color, their color is changed to a defined replacement color, and the cell number is added to a "queue" for further searches of neighbors. Once a cell has been checked, its position is added to a list of completed cells. This algorithm is referred to as "Four-way flood fill using a queue for storage".
Here's a visualization of the Four-way flood fill from Wikimedia Commons:
 |
| http://commons.wikimedia.org/wiki/File:Wfm_floodfill_animation_stack.gif |
This is kind of a pointless exercise given that any basic image editing programs (e.g. Microsoft Paint) can do this much more efficiently; Nevertheless, I felt compelled to figure out a way of programming this in R (I was originally interested in filling in land areas on a map that I created in R). You'll see from my example above that I didn't quite get it right - there is still some blank white space within the regions that I filled. Part of this problem is remedied by exporting a higher resolution image (floodfill argument "res"), but this slows things down considerably.
In order to have this function work directly on an open graphics device, I exported a PNG image and then re-imported it and trimmed off the margins. What remains is an image of the plot region itself which I convert to a matrix and look-up dataframe, where each cell's color and neighboring cells are defined. It is this dataframe that forms the basis of my searching algorithm. I'm guessing I have made some sort of small mistake in how I trimmed the margins of the image, thus creating the slight offset in the filled region. Anyway, feel free to suggest improvements!
Function:
