cv2.findCirclesGrid in python not working

Question

I want to calculate a calibration for my camera and I have this pattern but it for some reason doesnt recognise it.

calibration_pattern = cv2.imread('circles/IMG_9341.JPG')
gray_pattern = cv2.cvtColor(calibration_pattern, cv2.COLOR_BGR2GRAY)

# Define chessboard size and criteria
pattern_size = (4, 11)  # Change accordingly based on your calibration pattern

# Find circle centers on the calibration pattern
ret, corners = cv2.findCirclesGrid(gray_pattern, pattern_size, flags = cv2.CALIB_CB_ASYMMETRIC_GRID + cv2.CALIB_CB_CLUSTERING)

ret is always False

example image

I tried changing up the patterns, flags even other images but it is always False.

Answer 1

Your fix would be to edit the blobDetector params. The default blobDetector does not suit your image

Your circles are quite big, and hence if you set the minimum and maximum area of the blob detector, everything should work fine. Here's how to do it:

im = cv2.cvtColor(cv2.imread("circles.jpg"), cv2.COLOR_BGR2RGB) # read im
im_chess = im.copy()
gray = cv2.cvtColor(im, cv2.COLOR_RGB2GRAY) # convert to gray
blobParams = cv2.SimpleBlobDetector_Params() # create a default blob detector param
blobParams.filterByArea = True # enable filtering by area
blobParams.minArea = 10000   # minimum area
blobParams.maxArea = 40000   # maximum area
blobDetector = cv2.SimpleBlobDetector_create(blobParams) # create the blob detector
ret, corners = cv2.findCirclesGrid(im, (4,11),
                                   flags = (cv2.CALIB_CB_ASYMMETRIC_GRID+cv2.CALIB_CB_CLUSTERING),
                                   blobDetector = blobDetector) # pass the blobDetector to the function
if ret == True: # if detected...
    corners2 = cv2.cornerSubPix(gray, corners, (11,11), (-1,-1), 
                                (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 30, 0.001)) # refine corner location
    cv2.drawChessboardCorners(im_chess, (4,11), corners2, ret) # draw the chessboard

This is the end result:

Try to play around with other parameters if you want. This is a nice tutorial.