JPH066603A - Binarizing processing unit - Google Patents

Abstract

PURPOSE:To surely recognize a picture regardless of density. CONSTITUTION:A density signal subject to photoelectric conversion by a CCD 3 is fed to an average density calculation section 7 one by one picture element via an A/D converter 4. Latches 5, 6 delay the density signal one by one clock each and output the result to the average density calculation section 7. The section 7 calculates an average density A(a2) from each of density signals a1, a2, a3 sent sequentially. A correction section 9 adds a prescribed value alpha to the mean density A to obtain a threshold level S. A binarizing processing section 11 applies binarization processing to a density signal a2 at the threshold level S. Thus, the threshold level S is changed based on the density signal of picture elements adjacent to the noted picture element.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a binarization processing apparatus and, more particularly, to grayscale processing.

[0002]

2. Description of the Related Art FIG. 6 shows a conventional binarization processing device 51. The binarization processing device 51 reads an image of a read document with 64 gradations by the line CCD 3, compares the read density signal with a preset threshold value (fixed value), and binarizes each pixel. Is a device for performing.

A method of binarizing the read image by the binarization processing device 51 will be described with reference to FIG. When the image 16 shown in FIG. 9A is read by the reading line L1, the density signal G1 shown in FIG. 9B can be obtained. The binarization processing unit 53 (see FIG. 6) compares the density signal G1 with a predetermined threshold value S1. The binarization process is performed with white when the density signal G1 is higher than the threshold value S1 and black when the density signal G1 is lower than the threshold value S1. In this case, a binarized signal as shown in FIG. 7C is output. The image 16 can be binarized by shifting the line 16 in the sub-scanning direction F line by line and performing such a process.

[0004]

However, the binarization processing device 51 as described above has the following problems. As shown in FIG. 8A, it is assumed that there is an image 17 having a region V having a lighter density than the character H around the character H and an image 30 having the character H drawn with the same density as the region V. When this is read on the line L1, the density signal G2 as shown in FIG. This is the threshold S1
By binarizing with, the binarized signal as shown in FIG. In this case, the image 30 can be recognized as indicated by β, but the character H of the image 17 cannot be recognized as indicated by α.

On the other hand, in order to solve such a problem, if a predetermined threshold value is lowered and binarization processing is performed with the threshold value S2, a binarized signal as shown in FIG. 6C can be obtained. . In this case, the character H of the image 17 can be recognized as indicated by γ, but the image 30 cannot be recognized as indicated by δ.

An object of the present invention is to solve the above problems and to provide a binarization processing apparatus capable of surely recognizing an image regardless of the density.

[0007]

A binarization processing apparatus according to a first aspect of the present invention receives a density signal of an image sent from a photoelectric conversion means, and outputs a binary value for each pixel based on a predetermined threshold value. In the apparatus for performing the conversion processing, a threshold value setting means for changing the threshold value based on the density signals of pixels on both sides of each pixel is provided.

[0008]

The threshold value setting means changes the threshold value for the binarization processing based on the density signals of the pixels on both sides of each pixel. Appropriate binarization information can be obtained by binarizing the density signal of each pixel based on the obtained threshold value.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described with reference to the drawings. FIG. 1 shows a binarization processing device 1 according to the present invention.
The binarization processing device 1 includes a CCD 3 which is a photoelectric conversion unit, an analog / digital converter (A / D converter) 4, latches 5 and 6, an average density calculation unit 7, a correction unit 9, and a binarization processing unit 11. Is equipped with. In this example,
The latches 5 and 6, the average density calculation unit 7, and the correction unit 9 constitute a threshold value setting unit.

The binarization processing of the binarization processing apparatus 1 will be described. The density signal photoelectrically converted by the CCD 3 is input to the A / D converter 4 pixel by pixel according to a clock (not shown) and converted into digital data of multiple gradations (for example, 64 gradations). This digital data is sent to the average density calculator 7. Here, the latches 5 and 6 delay the density signal by one clock and output the delayed density signal to the average density calculator 7.

For example, when the density signal [a1] is sent from the CCD 3, the density signal [a1] is sent to the average density calculator 7
And sent to latch 5. The latch 5 uses the density signal [a
1] is temporarily stored. When the density signal [a2] is sent from the CCD 3 at the next clock, the density signal [a2]
Is sent to the average density calculator 7 and the latch 5. The latch 5 outputs the previously stored density signal [a1]. The outputted dark density signal [a1] is sent to the average density calculator 7 and the latch 6. The latch 6 temporarily stores the density signal [a1].

When the density signal [a3] is sent from the CCD 3 at the next clock, the density signal [a3] is sent to the average density calculation section 7 and the latch 5. The latch 5 outputs the previously stored density signal [a2]. The output dark density signal [a2] is output to the average density calculator 7 and the latch 6
Sent to. The latch 6 uses the density signal [a
1] is output to the average density calculator 7 and the density signal [a2] is stored.

At this stage, the average density calculator 7 is supplied with the density signals [a1], [a2], and [a3]. Here, assuming that the values of the density signals of the density signals [a1], [a2], and [a3] are a1, a2, and a3, the following equation (1) is obtained.
Thus, the average density A (a2) when the pixel that outputs the density signal [a2] is the target pixel is calculated.

A (a2) = (a1 + a2 + a3) / 3 (1) As described above, the latches 5 and 6 delay the density signal by one clock, and output the delayed density signal to the average density calculator 7. The average density A (a2) can be calculated.

The calculated average density A (a2) is added with a predetermined value α in the correction unit 9 in order to prevent it from being inverted and binarized. Sent to 11. That is, S is given by the following equation.

S = A (a2) + α (2) The density signal [a2] output from the latch 5 is sent to the binarization processing unit 11. The binarization processing unit 11 binarizes the value a2 of the density signal with the threshold value S.

In this embodiment, when the value a2 of the density signal is larger than the threshold value S (a2> S), the pixel is white. On the other hand, the value a2 of the density signal is the threshold value S
If (a2 ≦ S), the pixel is black.

In this way, the processing when the pixel a2 is the target pixel is completed. After that, the target pixels a3 and a4
By performing the same processing as ..., The image of the document can be binarized.

FIG. 2A shows an image 20 having an area V2 around the character I which is lighter in density than the character I and an image 22 having the same density as the area V2. Lines L3 of images 20 and 22
When read by, the density signal G2 as shown in FIG. In this case, the threshold value S changes based on the density signals of pixels on both sides of the pixel of interest as shown in FIG. Therefore, when binarized with the threshold value S, a binarized signal as shown in FIG. As a result, the image 20 having the region V2 around the character I which is lighter in density than the character I
Also, the image 22 in which characters are drawn with the same density as the area V2,
It can be binarized appropriately.

By the way, in the correction unit 9, the above (2)
The correction process like the formula is performed. This correction process will be described with reference to FIG. When the density signals as shown in A of the figure are sequentially sent, the density signal G5 as shown in B of the figure is obtained. The following results are obtained by calculating the average densities A (a102) to A (a106) of each pixel by the above equation (1).

A (a102) = (61 + 59 + 58) /3≈59.3 A (a103) = (59 + 58 + 56) /3≈57.7 A (a104) = (58 + 56 + 58) /3=57.3 A (a105) = (56 + 58 + 59) /3≈57.7 A (a106) = (58 + 59 + 61) /3≈59.3 Then, the obtained average densities A (a102) to A (a106) are used as threshold values to binarize each pixel. Then, as shown in FIG. 6C, the pixel a102 is black, the pixel a103 is white, the pixel a104 is black,
The pixel a105 is binarized such that the pixel a105 is white and the pixel a106 is black. That is, inversion of black and white occurs complicatedly. Therefore, in order to prevent such inversion binarization, the correction unit 9 adds the predetermined value α, and the value is set to the threshold value S.
I am trying. By performing such a correction, it is possible to prevent inversion binarization due to a minute fluctuation of the density value.

By the way, when the image 16 as shown in FIG. 3A is read by the reading line L3, the density signal G3 as shown in FIG. 3B can be obtained. When reading a portion where black is continuous in this way, the threshold value for the pixel in the central portion is
It becomes lower due to the density signals at both ends. Therefore,
The pixel in the center portion is binarized in white, and the image is recognized as shown in FIG. Therefore, in the present embodiment, a predetermined predetermined black limit value SLB is given to the binarization processing unit 11, and when the density signal of the pixel of interest is lower than the black limit value SLB, the threshold value S Regardless, it always binarizes as black. By performing such processing, it is possible to appropriately binarize the image 16 as shown in FIG.

Further, as shown in FIG. 4A, when a non-image having fine density unevenness 33 is read by the read line L4, a density signal G4 as shown in FIG. 4B is obtained. As described above, when a white-colored original having uneven density 33 is scanned, the pixels in the uneven density 33 are black, although it is originally necessary to binarize all white. Will be done.

Therefore, in this embodiment, the predetermined white limit value SLW is given to the binarization processing section 11,
When the density signal of the target pixel is higher than the white limit value SLW, it is always binarized as white regardless of the threshold value S (see D in the same figure). By performing such processing, even a white background original having uneven density as shown in FIG. 9A can be appropriately binarized.

Further, in this embodiment, the latches 5 and 6 are used.
Thus, the density signal is delayed by one clock and output to the average density calculation unit 7. However, it is also possible to use the line memory and temporarily store it in the line memory and send the values of the density signals of the three pixels to the average density calculation unit 7 at once.

[0026]

The binarization processing apparatus according to claim 1 receives the density signal of the image sent from the photoelectric conversion means,
A device for performing binarization processing for each pixel based on a predetermined threshold value is provided with threshold value setting means for changing the threshold value based on density signals of pixels on both sides of each pixel. And Therefore, by binarizing the density signal of each pixel based on the obtained threshold value, the image can be surely recognized regardless of the density.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a binarization processing device 1 which is an embodiment of the present invention.

FIG. 2 is a diagram showing a binarization processing method of the binarization processing apparatus 1.

FIG. 3 is a diagram showing a processing method when a black limit value SLB is provided in the binarization processing method.

FIG. 4 is a diagram showing a processing method when a white limit value SLW is provided in the binarization processing method.

FIG. 5 is an explanatory diagram of inversion binarization that occurs due to minute fluctuations in density value.

FIG. 6 is a block diagram showing a configuration of a conventional binarization processing device 51.

7 is a diagram showing a binarization processing method of the binarization processing device 51. FIG.

8 is a diagram showing a binarization processing method of the binarization processing device 51. FIG.

[Explanation of symbols]

 3 ... CCD 5, 6 ... Latch 7 ... Average density calculation unit 9 ... Correction unit 11 ... Binarization processing unit

Claims (1)

[Claims] 1. An apparatus for receiving a density signal of an image sent from a photoelectric conversion means and performing a binarization process for each pixel based on a predetermined threshold value, in a density signal of pixels on both sides of each pixel. A threshold value setting means for changing the threshold value based on the above.