Thanks for your kindness and help. And What's the meaning of "Meybe the SiSo support can provide a good hint."?
Another question: We got the CenterX & CenterY of all the objects After the blob analysis. If we want to combine two or more closed objects. Do you think is it possible to realize it through comparing the D-value of each CX & CY of all the objects？ How to get the D-value of each pixel of one line?