| 摘要 |
PURPOSE:To extract only a local peak point with good S/N by scanning a two- dimension histogram, extracting a peak candidate point having a prescribed level or over, classifying the point into classes and extracting the maximum level of the peak candidate point in the class. CONSTITUTION:A data Pi,j of the two-dimension histogram obtained by the Hough conversion is stored in a memory 1-1. A threshold arithmetic section 2 extracts a peak candidate point Xi having a prescribed threshold level or over from the data Pi,j and stores the coordinate and level to a memory 1-2. A cluster ring arithmetic section 3 compares the level of the Xi, rearranged into Yi in the larger order, operates Euclid interval with other points sequentially, compares the interval with the threshold so as to give labelling and classify the class. A maximum value operating section 4 references a cluster ring result of a memory 1-3 outputted from the cluster ring operating section 3 so as to obtain a maximum value Zi of each class.
|
