JPS63181521A - Floating binarization system - Google Patents

Abstract

PURPOSE:To precisely execute a binarization without receiving the effects such as the unevenness of illumination and the strain of shading, etc., by setting a floating binarization threshold following the waveform in an original signal in the case of binarization of a character pattern and a wiring pattern, etc. CONSTITUTION:When first order differentiation is executed to the original signal (a) so as to compute the absolute value of a differential value, a differential signal (c) is obtained and the projecting part of the waveform in the differential signal (c) is equal to the edge part of the waveform in the original signal (a). A threshold process is executed to the differential signal (c) with a proper differential threshold (d) in order to extract the edge part and a differential binarization signal (e) can be obtained. When the time, obtained by extracting the center of the high level period of the differential binarization signal (e), is t1, the time t1 is equal to the center part of the edge part in the original signal (a). As for a video signal including the strain obtained by image-picking up the character pattern and the wiring pattern, etc., only signal part can be precisely binarized.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention relates to a floating binarization method for accurately binarizing signals obtained from an imaging device such as a TV camera.

(Background technology) When a character pattern, wiring pattern, etc. shown in Figure 3 (a) is captured as an image by an imaging device such as a TV camera, ideally the video signal (for one scan) as shown in (b) is ) is obtained. Generally, due to the effects of uneven lighting, shading, etc., distortion is included as shown in (c). Signal waveforms are often obtained.

By the way, in the conventional binarization method, even if a fixed binarization threshold is used or a floating binarization threshold is adopted, it does not follow the original signal sufficiently, so it is difficult to distinguish objects from objects. There are drawbacks such as deterioration of binarization accuracy due to changes in the frequency component of the signal at the edge with respect to the background, and susceptibility to the effects of uneven illumination, shading, and the like.

(Object of the Invention) The present invention has been proposed in view of the above-mentioned points, and its purpose is to provide a signal processing method for video signals containing distortion obtained when imaging character patterns, wiring patterns, etc. The object of the present invention is to provide a floating binarization method that can binarize only the parts with high precision.

(Disclosure of invention) Hereinafter, the present invention will be explained in detail along with the drawings.

Figure 1 shows a signal waveform that undergoes processing and deforms using the floating binarization method of the present invention. Explain the steps to find.

First, the original signal a is first differentiated, and the absolute value of the differential value is calculated. At this time, a differential signal C is obtained. This differential signal C
The convex portion of the waveform corresponds to the edge portion of the waveform in the original signal a. In order to extract this edge portion, threshold processing is performed on the differential signal C using an appropriate differential threshold value d.

As a result, a differential binary signal e is obtained. The center of the high level period of this differential binary signal e is extracted, and this time is set as tl. Time t1 corresponds to the center of the edge portion in the original signal a.

This will be explained below using FIG. 2. Now, we convert the original signal a to the function y
=f(t). At this time, the values of points shifted by Δt before and after the point of interest t1 are f(tn+Δt) and f
It can be written as (tn-Δt). If the point where this binary value is divided into man is g(tnl, then g (t+) = (ni f
(i, +Δt)+n f (tn−Δt)1m+n. Similarly, time 10. Regarding 12, g(toL
g (calculate t21, point [:to, g (to)] and point [tl.

g (tn)), point Ctlp g(tn)) and point [t
2°g (t2)) are respectively defined as floating binarization thresholds. That is, it becomes floating binarization, and can generally be written as an interval tn-1≦1 (in 1°).

By thresholding the original signal a using this floating binarization threshold value b (=y), a floating binarization signal f is obtained.

(Effects of the Invention) As described above, in the present invention, a floating binarization threshold that follows the waveform of the original signal is set when character patterns, wiring patterns, etc. are binarized. Therefore, there is an effect that binarization can be performed with high precision without being affected by distortions such as uneven illumination and shading.

[Brief explanation of the drawing]

Figure 1 is a diagram showing a signal waveform that undergoes processing and deformation by the floating z-value conversion method of the present invention, Figure 2 is a diagram showing details of the floating binarization threshold, and Figure 3 is a diagram showing a captured image. and a diagram showing a video signal.

Claims (1)

[Claims] A step of firstly differentiating the original signal and calculating a differential signal indicating the absolute value of the differential value, and thresholding the differential signal using an appropriate differential threshold value to obtain a differential binary signal. a step of detecting a time t_n at the center of the high-level period of the differentiated binary signal; and a step of detecting a time t_n at the center of the high-level period of the differential binary signal; and a step of detecting a time t_n at the center of the high-level period of the differentiated binary signal; and assuming that the value of the original signal at time t is f(t), and a value of Δt before and after the detected time t_n; The values of the shifted points f(t_n+Δt), f(t
_n−Δt) and the above binary f(t_
n+Δt), f(t_n-Δt) divided into m:n to obtain a value g(t_n), and t_n_-_1≦t
< In the interval t_n, point [t_n_-_1, g(t_n_-
_1)] and the point [t_n, g(t_n)] y=(g(t_n_-_1)-g(t_n))/(t_
n_-_1-t_n) {x-g(t_n_-_1)}+
and the step of determining t_n_−_1, and the straight line y
A floating binarization method characterized in that the original signal is binarized using as a floating binarization threshold.