JP3019518B2 - Image reading device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reader for converting color digital image signals R, G, B signals into binary signals.

[0002]

2. Description of the Related Art Next, the configuration of a conventional image reading apparatus will be described with reference to FIG. 2 is an object, 6 is an image input unit, 7 is a comparison reference signal generation circuit, and 8 is a binarization determination circuit. For example, the object 1 is a printed wiring board, and the image input unit 6 is a camera. If the pattern of the printed wiring board is photographed with a camera, the reflectance is different between the pattern and the pattern other than the pattern. Therefore, the pattern of the printed wiring board and the pattern other than the pattern can be read from the output of the image input unit 6.

The image of the object 1 is converted into a digital image signal 21 by the image input unit 6. Comparison reference signal creation circuit 7
Calculates the comparison reference signal 22 from the density vs. frequency histogram of the image signal in advance. The method of calculating the comparison reference signal 22 is described in, for example, Nobuyuki Otsu, “Automatic threshold value calculation method based on discrimination and least square criterion,” IEICE Transactions on Electronics
1. Described in J63-D〓4. In this method, a threshold value that maximizes the degree of separation of a class to be separated is obtained from a density versus frequency histogram of an image signal.

Each time the image signal 21 is output from the image input unit 6, the binarization determination circuit 8 compares the image signal 21 with the comparison reference signal 22,
Value. The conditions for binarization are as follows. Image signal 21
> In the case of the comparison reference signal 22, the binary image signal 23 is set to a high level. 2 when image signal 21 ≦ comparison reference signal 22
The value image signal 23 is set to a low level.

Next, a graph of the pixel density versus the frequency of the number of pixels in the pattern portion and the portion other than the pattern when the object 1 is a printed wiring board will be described with reference to FIG. The horizontal axis in FIG. 4 is the pixel density, and the vertical axis is the frequency of the number of pixels. An overlap portion C exists in the density vs. frequency histogram A of the pattern portion and the density vs. frequency histogram B of the portion other than the pattern.

Next, the image of FIG.
The state when the value is converted will be described with reference to FIG. The horizontal axis in FIG. 5 is the density, and the vertical axis is the frequency. AB in FIG. 5 is a histogram of density vs. frequency of the image, and curves A and B in FIG.
Are synthesized. CA and CB in FIG. 5 correspond to the overlapping portion C in FIG. The overlapping portion CB in FIG. 5 is determined to be a portion other than the pattern, and the overlapping portion CA is determined to be a pattern and is binarized.

[0007]

It is difficult to correctly binarize an image signal having an overlapping portion C as shown in FIG. 4 using a density threshold. According to the present invention, parameters of lightness, saturation, and hue are obtained from a color image signal of the object 1, and these parameters are compared with a predetermined comparison reference signal to be binarized. It is an object of the present invention to provide an image reading device that can binarize an image signal.

[0008]

In order to achieve this object, according to the present invention, there is provided an image input unit for picking up an image of an object 1 and outputting an R signal 11, a G signal 12, and a B signal 13 of color digital image signals. 2, R signal 11, G signal 12, and B signal 1
3 as an input and outputs a color information separating circuit 3 for converting into a lightness signal 14, a chroma signal 15 and a hue signal 16, and a lightness comparison reference signal 17, a saturation comparison reference signal 18 and a hue comparison reference signal 19 Comparison reference signal generator 4 and brightness signal 1
4> output a high-level binary image signal 20 when lightness comparison reference signal 17 and saturation signal 15> saturation comparison reference signal 18 and hue signal 16> hue comparison reference signal 19; In other cases, a binarization determination circuit 5 that outputs a low-level binary image signal 20 is provided.

Next, the configuration of an image reading apparatus according to the present invention will be described with reference to FIG. 1 is an image input unit, 3 is a color information separation circuit, 4 is a comparison reference signal generator, and 5 is a binarization determination circuit. The image input unit 2 is, for example, a video camera or an image scanner. The image input unit 2 captures an image of the object 1, and outputs an R signal 11 and a G signal 1 of a color digital image signal.
2, the B signal 13 is output, and the color information separation circuit 3 converts the R signal 11, the G signal 12, and the B signal 13 from the image input unit 2 into a brightness signal 14, a saturation signal 15, and a hue signal 16.

[0010]

Next, the operation of the color information separation circuit 3 will be described. From the R signal 11, the G signal 12, and the B signal 13 output from the image input unit 2, the brightness signal 14, the saturation signal 15, and the hue signal 16 are obtained by the following equations. Brightness signal 14 = max (R, G, B) (1) Saturation signal 15 = ｛max (R, G, B) −min (R,
G, B)｝ / max (R, G, B) (2) When the hue signal 16 max (R, G, B) = R signal 11, [｛max (R, G, B) Signal B｝ / ｛max (R,
G, B) -min (R, G, B)｝]-[｛max
(R, G, B) -signal G / max (R, G, B)-
min (R, G, B)} + “1” When max (R, G, B) = G signal 12, [{max (R, G, B) −signal R} / {max (R,
G, B) -min (R, G, B)｝]-[｛max
(R, G, B) -Signal B｝ / ｛max (R, G, B)-
min (R, G, B)} + “3” When max (R, G, B) = B signal 13, [{max (R, G, B) −signal G} / {max (R,
G, B) -min (R, G, B)｝]-[｛max
(R, G, B) -signal R｝ / ｛max (R, G, B)-
min (R, G, B)｝] + “5” …………………
(3) where max (R, G, B) = max (R signal 11,
G signal 12, B signal 13), min (R, G, B) = m
in (R signal 11, G signal 12, B signal 13).

For the lightness signal 14, the saturation signal 15, and the hue signal 16, the lightness comparison reference signal value, the saturation comparison reference signal value, and the hue comparison reference signal value are set in advance in the comparison reference signal generator 4. Then, the comparison reference signal 17 for brightness, the comparison reference signal 18 for saturation, and the comparison reference signal 19 for hue are sent from the comparison reference signal generator 4 to the binarization determination circuit 5. The binarization determination circuit 5 includes a lightness signal 14 from the color information separation circuit 3,
A saturation signal 15 and a hue signal 16 are input.

In the binarization determination circuit 5, the brightness signal 14 is larger than the brightness comparison reference signal 17, the saturation signal 15 is larger than the saturation comparison reference signal 18, and the hue signal 16 is the hue comparison reference signal. It is determined whether it is greater than 19. If these are satisfied at the same time, the level is set to a high level; otherwise, the level is set to a low level.

Next, the configuration of the color information separation circuit 3 will be described. 3A to 3C of the color information separation circuit 3 are comparators, 3D is a subtractor, 3E is a divider, 3F and 3G are selectors, 3H and 3J.
Is a subtractor, 3K and 3L are dividers, 3M is a subtractor, 3N is a selector, and 3P is an adder. The comparator 3A has an R signal 1
1 and the G signal 12 are input, and the comparator 3A sets the larger one of the R signal 11 and the G signal 12 to the signal 31 and sets the smaller one to the signal 32. The signal 31 and the B signal 13 are input to the comparator 3B. The comparator 3B sets the larger one of the signal 31 and the B signal 13 as the brightness signal 14. The signal 32 and the B signal 13 are input to the comparator 3C.
The smaller one of the three signals is signal 33.

The brightness signal 14 is input to a of the subtracter 3D, and the signal 33 is input to b. The subtractor 3D is a-
The result of b is output as a signal 34. The signal 34 is input to a of the divider 3E, and the brightness signal 14 is input to b. The divider 3E outputs the result of a / b as the saturation signal 15.

The selector 3F includes an R signal 11, a G signal 12,
The B signal 13 and the brightness signal 14 are input, and the signal 35 is extracted. In this case, when the brightness signal 14 = R signal 11, the B signal 13 becomes the signal 35, when the brightness signal 14 = G signal 12, the R signal 11 becomes the signal 35, and when the brightness signal 14 = B signal 13, the G signal becomes The signal 12 becomes the signal 35.

The selector 3G includes an R signal 11, a G signal 1
2, the B signal 13 and the brightness signal 14 are input, and the signal 36
Is taken out. In this case, the brightness signal 14 = the R signal 11
, The G signal 12 becomes the signal 36, and the brightness signal 14 =
When the G signal 12, the B signal 13 becomes the signal 36, and when the brightness signal 14 = B signal 13, the R signal 11 becomes the signal 36.

The brightness signal 14 is input to a of the subtractor 3H, and the signal 35 is input to b. Subtractor 3H is ab
And outputs the result as a signal 37. The lightness signal 14 is input to a of the subtractor 3J, and the signal 36 is input to b. The subtractor 3J calculates ab and outputs the result as a signal 38.

A signal 37 is input to a of the divider 3K.
The signal 34 is input to b. The divider 3K calculates a / b and outputs the result as a signal 39. The signal 38 is input to a of the divider 3L, and the signal 34 is input to b. The divider 3L calculates a / b and outputs the result as a signal 40. The signal 39 is input to a of the subtractor 3M, and the signal 40 is input to b.
Is entered. The subtractor 3M calculates a−b, and outputs the signal 41
Output as

The selector 3N includes an R signal 11, a G signal 1
2, the B signal 13 and the brightness signal 14 are input, and the signal 42
Is taken out. In this case, the brightness signal 14 = the R signal 11
In this case, "1" is output, when brightness signal 14 = G signal 12, "3" is output, and when brightness signal 14 = B signal 13, "3" is output. The signal 41 is input to a of the adder 3P, and the signal 42 is input to b. The adder 3P is a +
b is calculated and output as the hue signal 16.

Next, the configuration of the binarization determination circuit will be described with reference to FIG. 5A, 5B and 5C are comparators and 5D is A in FIG.
ND gate. The comparator 5A outputs the lightness signal 14 to a
And the brightness comparison reference signal 17 is input to b. When a> b, the signal 51 goes high, and a ≦ b
At this time, the signal 51 goes low. The comparator 5B includes
A saturation signal 15 is input to a, and a saturation comparison reference signal 18
Is input to b. The signal 52 goes high when a> b, and goes low when a ≦ b. To the comparator 5C, the hue signal 16 is input to a, and the hue comparison reference signal 19 is input to b. The signal 53 goes high when a> b, and goes low when a ≦ b. The signal 51, the signal 52, and the signal 53 are AND
The binary image signal 20 is output through the gate 5D.

Next, the distribution of lightness, saturation and hue will be described with reference to FIG. 3 using a printed wiring board as an example. 3A shows the distribution of the pattern portion of the printed wiring board, and BB shows the distribution of the portion other than the pattern. In the case where AA and BB in FIG. 3 are binarized, if the value is larger than the lightness comparison reference signal 17, larger than the saturation comparison reference signal 18, and larger than the hue comparison reference signal 19, the BB portion Is low, A
A becomes high level, otherwise BB
Is at a high level and AA is at a low level. In this case, since AA and BB do not overlap, binarization can be performed correctly.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The value of the R signal 11 in FIG.
Is "40" and the value of the B signal 13 is "30", the signal 31 is "40", the signal 32 is "20", and the brightness signal 14
Becomes "40". Further, the signal 33 becomes “20”, and the subtractor 3D subtracts “40” − “20” = “20”. The divider 3E divides “20” ÷ “40” = “0.5”, and the saturation signal 15 becomes “0.5”.

The selector 3F has "20" for a and "4" for b.
“0”, “30” is input to c, and “40” is input to s. At this time, since the brightness signal 14 = the G signal 12, the signal 35 = R
The signal 11 becomes "20". "20", b for selector 3G
"40", "30" for c, and "40" for s. At this time, since the brightness signal 14 = the G signal 12, the signal 36 = the B signal 13 "30".

The subtractor 3H subtracts the brightness signal 14 “40” −the signal 35 “20” = the signal 37 “20”, and the subtractor 3J performs the brightness signal 14 “40” −the signal 36 “30” = the signal 38 ”. 1
Subtract "0". In the divider 3K, the signal 37 “20” ÷ the signal 34 “20” = the signal 39 “1” is divided, and the divider 3L is divided.
Then, the signal 38 “10” ÷ the signal 34 “20” = the signal 40 “0.
Divide 5 ”. The subtracter 3M subtracts the signal 39 “1” −the signal 40 “0.5” = the signal 41 “0.5”.

The selector 3N has "20" for a and "4" for b.
“0”, “30” is input to c, and “40” is input to s. At this time, since the brightness signal 14 = the G signal 12, the signal 42 becomes "1". The adder 3P adds the signal 41 "0.5" + the signal 42 "1" = the hue signal 16 "1.5". As a result, the value of the R signal 11 is “20” and the value of the G signal 12 is “4”.
0 ”and the value of the B signal 13 is“ 30 ”.
Is "40", the chroma signal is 0.5, and the hue signal 16 is "1.5".

In the binarization determination circuit 5 of FIG. 6, the value of the lightness signal 14 is "40", the value of the chroma signal 15 is "0.5", the value of the hue signal 16 is "1.5", and the comparison standard of the lightness. If the value of the signal 17 is “30”, the value of the saturation reference signal is “0.4”, and the value of the hue comparison reference signal is “1.0”, the comparator 5A
Since "40" is input to "a" and "30" is input to b, the signal 51 becomes high level. "0.5" is set for a of the comparator 5B.
However, “0.4” is input to b, the signal 52 becomes high level, “a” of the comparator 5C is “1.5”, and “b” is “1.0”.
Is input, the signal 53 goes high. All inputs of the AND gate 5D become high level, and the binary image signal 20
Is at a high level.

In FIG. 3, it is assumed that the distribution BB of the portion other than the pattern has a lightness of "120", a saturation of "0.3" and a hue of "2" in a spherical shape. On the other hand, it is assumed that the pattern distribution AA is spherically distributed with lightness “180”, saturation “0.45”, and hue “3”. At this time, in order to adjust the scales of the lightness axis, the saturation axis, and the hue axis, the saturation axis is assumed to be multiplied by 400, and the hue axis by 60 times.

The distribution range of the brightness of the distribution AA is "85" to "15".
5 ", the saturation distribution range is" 0.22 "to" 0.38 ", the hue distribution range is" 1.4 "to" 2.6 ", the lightness distribution range of the distribution BB is" 145 "to" 215 ", and the saturation distribution is Change the distribution range to "0.3
7 "to" 0.53 ", and the hue distribution range from" 2.4 "to" 3.6 "
And At this time, the brightness comparison reference signal 17 is set to “14”.
5 ", the saturation reference signal 18 is set to" 0.36 ", and the hue comparison reference signal 19 is set to" 2.38 ".

In this case, the maximum value of the brightness on the main diagonal axis is 12
0+ (155−120) / √3 = 140, maximum saturation is 0.3
+ (0.38−0.3) /√3=0.35, the maximum value of hue is 2+
(2.6−2) /√3=2.35, the minimum value of the brightness of the pattern portion is 180+ (145−180) / √3 = 160, and the minimum value of the saturation is 0.45+ (0.37−0.45) / √3 = 0.40, and the minimum value of the hue is 3+ (2.4−3) /√3=2.65. In this case, the distribution AA of the portion other than the pattern is at a low level, and the distribution BB of the pattern portion is at a high level. Since AA and BB do not overlap, binarization can be performed correctly.

[0030]

According to the present invention, a color image signal is extracted from an object, parameters of a lightness signal, a saturation signal, and a hue signal are obtained from the image signal, and the parameters are compared with a preset comparison reference signal. Are binarized by comparing with each other, it is possible to reduce binarization errors and binarize even an image signal having an overlap with the density vs. frequency graph.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an image reading apparatus according to the present invention.

FIG. 2 is a configuration diagram of a conventional image reading apparatus.

FIG. 3 is a distribution diagram of lightness, saturation, and hue taking a printed wiring board as an example.

FIG. 4 is a graph of pixel density versus frequency of a pattern portion and a portion other than the pattern.

5 is a state diagram when the image of FIG. 4 is binarized by a comparison reference signal 22. FIG.

FIG. 6 is a configuration diagram of a binarization determination determination circuit 5.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Object 2 Image input part 3 Color information separation circuit 4 Comparison reference signal generator 5 Binarization determination circuit 11 R signal 12 G signal 13 B signal 14 Lightness signal 15 Chroma signal 16 Hue signal 17 Lightness comparison reference signal 18 Saturation comparison reference signal 19 Hue comparison reference signal 20 Binary image signal

Claims (1)

(57) [Claims] An image input unit (2) for imaging an object (1) and outputting an R signal (11), a G signal (12), and a B signal (13) of a color digital image signal; 11) is compared with the G signal (12), and the larger signal is used as the first signal (31), and the smaller signal is used as the second signal (32).
The second comparator (3B) compares the B signal (13) with the first signal (31) and determines the larger one as the brightness signal (14).
And a third comparator (3C) that compares the B signal (13) with the second signal (32) and determines the smaller one as the third signal (33);
A first subtractor (3D) for subtracting the third signal (33) from (14)
A first divider (3E) that divides the output of the first subtractor (3D) by the brightness signal (14) to obtain a saturation signal (15); and an R signal (11);
G signal (12), B signal (13) and brightness signal (14) are input,
When brightness signal (14) = R signal (11), B signal (13) is output,
When brightness signal (14) = G signal (12), R signal (11) is output,
A first selector (3F) that outputs a G signal (12) when the brightness signal (14) = B signal (13), an R signal (11), a G signal (12), and a B signal
(13) and brightness signal (14) as input, output G signal (12) when brightness signal (14) = R signal (11), and B signal when brightness signal (14) = G signal (12) A second selector (3G) that outputs (13) and outputs an R signal (11) when the brightness signal (14) = B signal (13)
A second subtractor (3H) for subtracting the output of the first selector (3F) from the brightness signal (14); and a second selector from the brightness signal (14).
A third subtractor (3J) for subtracting the output of (3G), and a second subtractor for dividing the output of the second subtractor (3H) by the output of the first subtractor (3D).
Divider (3K), a third divider (3L) for dividing the output of the third subtractor (3J) by the output of the first subtractor (3D), and a second divider (3K) A fourth subtractor (3M) for subtracting the output of the third divider (3L) from the output of the third signal, an R signal (11), a G signal (12), and a B signal
(13) and the brightness signal (14) are input, and when the brightness signal (14) = R signal (11), "1" is output, and the brightness signal (14) = G signal (1
The output of the third selector (3N) and the fourth subtractor (3M) that output “3” when 2) and output “5” when the brightness signal (14) = B signal (13) And an adder (3P) that adds the output of the third selector (3N) to the hue signal (16).
A color information separation circuit (3) for converting the signal (12) and the B signal (13) into a brightness signal (14), a saturation signal (15) and a hue signal (16); a brightness comparison reference signal (17); A comparison reference signal generator that outputs a saturation comparison reference signal (18) and a hue comparison reference signal (19)
(4), and outputs a high level when the brightness signal (14)> the brightness comparison reference signal (17) and outputs a low level when the brightness signal (14) ≦ the brightness comparison reference signal (17) The fourth comparator (5A) outputs a high level signal when the saturation signal (15)> the saturation reference signal (18), and outputs the saturation signal (15) ≦ the saturation reference signal (18). )), The fifth comparator (5B) that outputs a low level, and the hue signal (16)
> Output high level when the hue comparison reference signal (19)
When the hue signal (16) ≦ the hue comparison reference signal (19), the sixth comparator (5C) that outputs a low level, the output of the fourth comparator (5A), and the fifth comparator (5B) ) And an output of a sixth comparator (5C), and a binarization determination circuit (5) that outputs a binary image signal (20). A featured image reading device.