Wsrtjtyk

I have a very simple usecase, I am gridding an image and calculating the most frequent occurring colour in each cell.

I have a problem where certain cells that visually look like black should be the most frequent colour have a mode of white (255). An example of this is cell (6,3) - zero-based index column first then row.

It should be black shouldn't it but it turns out white. Is there an issue with my code or is there salt and pepper noise I cannot see?

def mode_filter(self, roi):

    values = np.zeros((1, 256), dtype="uint8")



    for pos, val in np.ndenumerate(roi):

        values[0, val] += 1



    print(values)

    return int(np.argmax(values[0]))



def get_roi(self, src, pt1, pt2):



    col1, col2 = (pt1[0], pt2[0]) if pt1[0] < pt2[0] else (pt2[0], pt1[0])

    row1, row2 = (pt1[1], pt2[1]) if pt1[1] < pt2[1] else (pt2[1], pt1[1])

    return src[row1:row2, col1:col2]



def grid_img(self, src, nCols=7, nRows=7):



    gray = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)

    cells = np.zeros(gray.shape, dtype="uint8")

    cell_w = int(gray.shape[1] / nCols)

    cell_h = int(gray.shape[0] / nRows)



    for c in range(nCols): 

        for r in range(nRows):

            roi = self.get_roi(gray, (c*cell_w, r*cell_h), ((c+1)*cell_w, (r+1)*cell_h))



            mode = self.mode_filter(roi)

            cv2.rectangle(cells, (c*cell_w, r*cell_h), ((c+1)*cell_w, (r+1)*cell_h), (mode,), -1)



            cv2.imshow('roi', roi)

            cv2.imshow('cells', cells)

            cv2.imshow('src', src)

            print('{}, {}'.format((c,r), mode))

            cv2.waitKey(0)



    return cells

A full working example can be found here: https://pastebin.com/k7kUTmXG

The definition of the array you use to count the number of occurrences of each colour is as follows:

values = np.zeros((1, 256), dtype="uint8")

That means that each count is represented by a single byte, meaning the maximum value is 255. Once an additional pixel of that colour appears, the count will overflow to 0.

This is pretty much guaranteed to happen, as the size of the ROI you're processing is ~ 100x100 pixels.

The solution is simple -- use a bigger data type, such as int32.

def mode_filter(self, roi):

    values = np.zeros((1, 256), dtype="int32")



    for pos, val in np.ndenumerate(roi):

        values[0, val] += 1



    print(values)

    return int(np.argmax(values[0]))

For example, on tile (0,5), which looks like this:

your function gives the following output:

[[144  15   9  13   8   1   5   9   4   5   5   1   2   4   4   4   4   6

    7   5   4   2   1   2   1   1   0   1   2   1   6   3   4   7   2   3

    1   1   2   1   0   4   2   2   3   2   0   2   0   2   0   1   2   5

    1   3   2   4   2   2   3   3   1   3   4   2   2   2   2   4   3   2

    4   0   1   2   0   2   2   2   1   0   2   2   1   1   2   1   0   3

    2   0   1   1   0   4   1   3   1   2   3   1   0   3   0   1   1   0

    0   1   2   1   2   0   1   2   2   1   1   0   1   2   2   4   2   1

    2   1   1   1   1   2   1   1   1   3   1   1   2   3   2   1   2   0

    1   1   1   3   4   2   3   1   1   2   1   1   5   3   1   2   1   1

    1   0   2   3   0   1   1   3   4   0   2   1   4   3   0   1   4   1

    1   2   2   1   0   1   1   2   1   1   1   3   2   3   1   5   0   2

    2   1   2   2   0   2   1   0   1   1   4   2   4   2   2   4   1   4

    4   4   1   3   1   0   3   6   3   2   0   2   3   5   3   3   5   5

    5   4   3   6   7   4   1   7   2  10   5  10   6  10   7   7  16   8

   10  14  16 125]]

0

whereas the corrected version produces this:

[[2960   15    9   13    8    1    5    9    4    5    5    1    2    4

     4    4    4    6    7    5    4    2    1    2    1    1    0    1

     2    1    6    3    4    7    2    3    1    1    2    1    0    4

     2    2    3    2    0    2    0    2    0    1    2    5    1    3

     2    4    2    2    3    3    1    3    4    2    2    2    2    4

     3    2    4    0    1    2    0    2    2    2    1    0    2    2

     1    1    2    1    0    3    2    0    1    1    0    4    1    3

     1    2    3    1    0    3    0    1    1    0    0    1    2    1

     2    0    1    2    2    1    1    0    1    2    2    4    2    1

     2    1    1    1    1    2    1    1    1    3    1    1    2    3

     2    1    2    0    1    1    1    3    4    2    3    1    1    2

     1    1    5    3    1    2    1    1    1    0    2    3    0    1

     1    3    4    0    2    1    4    3    0    1    4    1    1    2

     2    1    0    1    1    2    1    1    1    3    2    3    1    5

     0    2    2    1    2    2    0    2    1    0    1    1    4    2

     4    2    2    4    1    4    4    4    1    3    1    0    3    6

     3    2    0    2    3    5    3    3    5    5    5    4    3    6

     7    4    1    7    2   10    5   10    6   10    7    7   16    8

    10   14   16 7293]]

255