| 摘要 |
PURPOSE:To recognize background density distribution from a picture in which a pattern is distributed arbitrarily by finding the respective density histograms of two pictures within window, defining the central density value of a low density side as background density and when a central value cannot be obtained, defining the central density value of the density histogram within window as the background density. CONSTITUTION:Density histograms 4 and 5 are calculated on two pictures within window respectively, a density differential histogram is calculated and when the sum of the absolute values of the histogram is '0' or smaller than an assigned value, it is decided whether it is a pattern part or not from the average of the histogram within window. When the result of it is not the pattern part, the peak or the statistical average of it is calculated, the value is defined as background density, when the sum of the absolute values is not '0' or larger than the assigned value, the polar part of the low density side of the density differential histogram is fetched and the peak or the statistical average of it is calculated. Thus, background density distribution can be recognized from a picture in which a pattern is distributed arbitrarily. |
