| 摘要 |
PURPOSE:To extract a graphic form border line efficiently by generating a picture element state value that an aimed picture element has from the coordinate values of the aimed picture element on the border line and two picture elements which are before and behind the aimed picture element on the border line, detecting a tracking start candidate picture element previously, and marking only the picture element. CONSTITUTION:The aimed picture element is denoted as P, the picture element before it as P-, and the picture element behind it as P+. In this case, the direction CH1 from P- to P and the direction CH2 from P to PL are represented with Freeman chain codes (0-7). Then the picture element state value (0, 1) that the aimed picture element P has is determined in a table (a) according to the combination of the directions CH1 and CH2 and processing corresponding to this picture element state value is carried out. When the picture element state value is 1, a picture element value '1' is written at a corresponding position on a work memory and when the picture element value is '1', a tracking end picture element is decided and a next tracking start candidate picture element is searched for; when the picture element value is '0', an untracked picture element is decided and the border line begins to be tracked from this picture element. Consequently, the border line of a graphic form can be extracted efficiently. |
