Image Processing Reference
Fig. 13.6 The different segmentation steps. ( a ) The input image. ( b ) The input thresholded using
the global method. ( c ) The mean of the input image calculated with a 200 × 200 kernel. ( d )The
result of subtracting the mean from the input. ( e ) Image ( d ) thresholded using the global method.
( f ) The effect of applying morphologic closing with a kernel of 7 × 7 to the binary image in ( e )
“I started out using just the size, but that was not enough and I therefore ended
up also using the number of holes in the BLOB.”
“But how will you figure out if a BLOB contains a hole or not?”
“I'll find the center of the BLOB and see if that pixel is black or white,” Mick
“But what if the center pixel is black due to noise that is not removed in the
segmentation process?” Mick felt quiet. “Argh, didn't think about that”. Silence.