TW243515B - Image thinning method and device - Google Patents

Abstract

An image thinning method with one pass parallel process, which builds the template matched with the input image in the template matching device, includes the following steps: 1. inputting one image data and storing the data in one image buffer; 2. clearing the flip-flop to LOW, which is used to decide if splitting the edge points; 3. inputting the neighboring point of each pixel into each processing element of template matching processor from image buffer, proceeding template matching, if the neighboring point of one black pixel is matched with any template built in the template matching processor, then clearing the black pixel to white, storing it in image buffer, and setting the flip-flop to HIGH; 4. repeating step 2 and 3 until the flip-flop is LOW; in which the templates includes the following 14 one (a) 1 1 y 1 c 0 1 1 y (b) 1 1 1 1 c 1 y 0 y (c) y 1 1 x 0 c 1 1 y 1 1 x (d) y 0 y 1 c 1 1 1 1 x 1 x (e) x 0 0 1 c 0 x 1 x (f) x 1 1 0 c 1 0 0 x (g) 0 1 0 0 c 1 0 0 0 (h) x 1 x 1 c 0 x 0 0 (i) 0 0 x 0 c 1 x 1 1 (j) 0 0 0 0 c 1 0 1 0 (k) 0 0 0 0 c 0 1 1 1 (l) 1 0 0 1 c 0 1 0 0 (m) 1 1 1 0 c 0 0 0 0 (n) 0 0 1 0 c 1 0 0 1 in which the symbol 'c represents testing point, symbol 'x' represents ignoring point, symbol '0' represents which must be one white point, symbol '1' represents which must be one black point and symbol 'y' must appear in pair representing there is at least one white point among the two pints.