原文:Next, to count the major blobs that are the keypads, we use the labeling algorithm. To handle a single keypad in an image, we proceed to separate the large blobs into separate images. In each individual image, we fill in the small black spots inside the big blobs and replace the pixel values of the white spots with the pixel values of the input image as seen in Fig. 1.11d–f. Then, we move the keypad in the middle of the image and try to rotate the image so it matches the orientation and position of the reference image.
Figure 1.11 (a) Input image; (b) resulting image after calibration; (c) binary image and detecting the big blobs; (d–f) resulting image after detecting the type of keypads; (g–i) resulting image after the keypads are moved and rotated horizontally.
斑点
这里的blobs就是黑色的小点点,可以看a图,如第一句所述,这些小黑点点主要就是一个个的按键