Technique of estimation of probability of recognition of text symbols on noisy images

Abstract

The paper presents a method of experimental evaluation of the probability of recognition of
text symbols (letters and numbers) on noisy images. The description of special software for experimental research is given. As a result of processing of results of experimental researches analytical relations for calculation of probability of the correct recognition by the operator of letters
and numbers (font - Times New Roman, Kegel - 14 points) displayed on the screen of the monitor, depending on the relation a signal/noise of the image are received. The ratio for calculating
the coefficient of improvement of the signal-to-noise ratio due to the accumulation (averaging)
of images is also obtained.