Question

I’m following this tutorial
http://aishack.in/tutorials/sudoku-grabber-with-opencv-detection/
But it’s not fot the iOS.
some methods described there can be used, but the floodfill (used as cvFloodFill) is no longer an int. it’s a void

I’m trying to get the size of the biggest blob, It’s pretty much the part i’m missing. this for loops and the floodfill.

is there a good way of doing it?

I thought about the cvBlob library, but I just can’t get it to compile for mac.. tried many ways….

EDIT: After arkiaz’s answer, moved to a new question:

openCV cvContourArea

Thanks!

Answer 1

Instead of floodfill you can use findcontour to find biggest blob.

Do as follows:

1) Apply adaptive threshold, some erosion and dilation etc ( as mentioned in aishack)

2) Find contours using findcontours

3) Use contourarea to find area of each contour

5) Select the one with max. area. That will be the sudoku box. Then continue as given in aishack.in

For me, this method worked faster than floodfill method