Binary conversion process for picture data - counting black pixels with threshold changed from darkest to lightest tone level for pattern recognition

摘要

A binary conversion process maintains an optimum threshold value for a binary conversion which is used when multi-level picture data, which draws an input picture, is converted into black/white picture data with two levels. This process counts up an initial number of black pixels, while a threshold value is changed from a very dark to a very bright tone level. It counts up a second number of pixels, which have a tone level outside the bright tone level. A proportion of the initial number is kept for each of the changed threshold values with regard to the second number, in order to standardise any density in the input picture. The optimum threshold value is set for a binary conversion on the basis of the proportion, which is kept for each of the changed threshold values.