JP2000276553A - Method for binarizing information code, optical information reader using this binarizing method, and code quality judging auxiliary equipment using this binarizing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make appropriately binarizable multivalued data without depending on the area ratio of an information code area in fetched image data to a background area. SOLUTION: First, a reference value is calculated (S100 to S120), the average gradation value (dark side average value) of 'dark' side pixels is calculated (S150) by designating gradation values being equal to or less than the reference value (S130) and executing designated range average value calculation processing (S140), the average gradation value (light side average value) of 'light' side pixels is also calculated (S180) by designating gradation values being equal to or more than the reference value (S160) and executing designated range average value calculation processing (S170), and a threshold is decided (S210) by repeating processing that makes the arithmetic mean a new boundary value (reference value) (S220).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technology for optically reading information codes such as bar codes and two-dimensional codes.

[0002]

2. Description of the Related Art Conventionally, an apparatus for reading an information code has been known. For example, in a conventional two-dimensional code reader, an analog signal read and output by a CCD sensor or the like is subjected to A / D conversion to obtain multi-value data (multi-tone digital image data) corresponding to the signal level. Data is stored once in an image memory, and the multi-valued data stored in the image memory is converted into binary data (binary digital image data) (hereinafter referred to as "binarization") and decoded. there were.

There are various types and sizes of two-dimensional codes. Particularly, when reading a high-density two-dimensional code, the number of pixels per cell as an information unit constituting the two-dimensional code is considered. This is because the number of (assigned pixels) decreases, and whether a pixel corresponding to a boundary portion of light (white) / dark (black) is determined as “bright” or “dark” greatly affects the success or failure of reading. .

[0004] As a method of binarizing multi-value data, AI is used.
According to the international standard created by M, all image pixel values are sorted by reflectivity and the average of the reflectivity containing the darkest 10% and the lightest 10% of the pixels is thresholded. , And the quality of the code is determined using the binarized image created by the binarization.

Here, this binarization method will be described in detail. First, if it is multi-valued data in which each pixel has a gradation value in the range of 0 to 255, the pixels are sorted by this gradation value, and 10% of the pixels are dark (the gradation value is dark). A range of the gradation value including the pixel on the (small) side is obtained. This range is indicated by gradation values “0” to “α”. Similarly, a range of gradation values that is 10% of the pixels and includes pixels on the brighter side (larger gradation values) is obtained. This range is indicated as gradation values “β” to “255”. At this time, the threshold is obtained as (α + β) / 2.

For example, a two-dimensional code image read in a visual field as shown in FIG. The shaded area in FIG. 10A indicates a two-dimensional code area, and this two-dimensional code area is assumed to be composed of light or dark cells. In this case, FIG. 11A schematically shows a distribution indicating the correspondence between the gradation value and the number of pixels. As described above, the distribution of the gradation value and the number of pixels is generally divided into two. That is, the distribution of the pixels corresponding to “dark” appears on the lower gradation value side, and the distribution of the pixels corresponding to “bright” appears on the higher gradation value side. In this case, as shown in FIG. 11A, 10% of the pixels on the dark side are included up to the gradation value α1, and 10% of the pixels on the light side are included in the portion having the gradation value β1 or more. include. Therefore, the threshold value γ1 is (α1 + β1)
/ 2. If the multi-valued data image is binarized using the threshold value γ1, pixels having a tone value smaller than the threshold value γ1 are regarded as “dark”, and conversely, pixels having a tone value greater than γ1 are regarded as “bright”. It will be binarized. As a result, the distribution of pixels corresponding to “dark” is distributed in a lower gradation value than the threshold value γ1, so that the distribution of pixels corresponding to “bright” is higher than the threshold value γ1. Since it is distributed in the higher gradation value, it is binarized as “bright”.

[0007]

However, in the case where this technique is used, for example, as shown in FIG. 10B, when the proportion of the two-dimensional code area in the whole image data becomes small, When the area of the background region increases, the following inconvenience occurs. In addition,
FIG. 10 shows the cells constituting the two-dimensional code area.
It is assumed that the background area is “bright” (white) as in (a).

At this time, FIG. 11B schematically shows a distribution indicating the correspondence between the gradation value and the number of pixels. FIG.
As can be seen from (b), the number of pixels corresponding to “dark” decreases and the number of pixels corresponding to “bright” increases because the number of pixels corresponding to “dark” in the image data is extremely small. are doing. Therefore, as shown in FIG. 11B, 10% of the pixels on the dark side are included up to the gradation value α2. That is, it is larger than the above-mentioned gradation value α1. On the other hand, 10% of the pixels on the bright side are included in the portion having the tone value β2 or more. Therefore, the threshold γ2 is (α
2 + β2) / 2, and the threshold γ2 is
It becomes larger than the threshold value γ1 shown in FIG. Then, if the multi-valued data image is binarized using the threshold value γ2, a pixel having a gradation value smaller than the threshold value γ2 is binarized as “dark”. Therefore, a pixel that should correspond to “bright” as indicated by the symbol A in FIG. 11A is binarized as “dark”. As a result, FIG.
In the case of (b), there is a high possibility that a pixel corresponding to a light / dark boundary portion of each cell constituting the two-dimensional code becomes “dark”, and when the binary data is restored as an image, each cell becomes “black”. The result is "fat", which adversely affects reading. When the quality of a two-dimensional code is determined using an image based on the binary data, FIG.
Even when the same two-dimensional code as in (a) is read,
It is determined that the quality is lower than that of FIG.

As a method for solving this problem, for example, the mode α3 of the distribution of pixels corresponding to “dark” and the mode of distribution of pixels corresponding to “bright” in FIG. β3
(Α3 + β3) / 2 is set as the threshold γ3. For example, when the area of the two-dimensional code image area becomes extremely small as compared with the background image area, for example,
The number of pixels corresponding to "dark" decreases,
If the gradation value of the pixel corresponding to “dark” varies, a situation may occur in which the mode α3 cannot be obtained.

An object of the present invention is to appropriately binarize multi-valued data regardless of the area ratio between an information code area and a background area in captured image data.

[0011]

According to a first aspect of the present invention, there is provided an information code binarizing method, wherein an image to be read is optically detected. For example, the reflected light from the object to be read is converted into an electric signal using a CCD sensor. Then, by subjecting the analog image signal corresponding to the detected image to A / D conversion, the image to be read is stored as multi-value data in which each pixel has a gradation value corresponding to the signal level. For example, each pixel is stored as multi-value data having an 8-bit gradation value. In this case, since there are 2 ⁸ (= 256) gradations, each pixel has a gradation value from “0” to “255”.

Then, the stored multi-value data is converted into binary data using the threshold value determined in the threshold value determination process. That is, each pixel of multi-gradation is converted into two gradations of “bright” or “dark”. In the present invention, the information code is binarized by such a basic procedure. The binarization method of the present invention is characterized by the above-described threshold value determination processing. Therefore, the threshold value determination processing will be described below.

Since the threshold value determination processing in the present invention includes the specified range average value calculation processing, the specified range average value calculation processing will be described first. The designated range average value calculation process is a process of calculating a lower gradation value and an upper gradation value, and calculating an arithmetic average of both gradation values.

The lower gradation value is a predetermined percentage of the pixels having the gradation value in the designated range, and is the largest of the gradation values of the pixels having the relatively small gradation value. A thing. Similarly, the upper gradation value is a predetermined percentage of the pixels having the gradation value in the specified range, and is the minimum gradation value of the pixel having the relatively large gradation value. A thing.

For example, in the example shown in FIG. 11, it is assumed that the gradation values "0" to "255" are designated and the predetermined ratio is set to 10% to execute the designated range average value calculation processing.
At this time, pixels that are 10% of all pixels and have relatively small gradation values are included in gradation values “0” to “α1” in FIG. In (b), it exists in “0” to “α2”. Therefore, the maximum gradation value of these pixels is “α” in FIG.
1 ”and“ α2 ”in FIG. 11B. Similarly, a pixel having a relatively large gradation value, which is 10% of all pixels, has a gradation value “β1” in FIG.
11 to “255”, and “β2” in FIG.
~ 255. Therefore, the minimum gradation value of these pixels is “β1” in FIG. 11A, and is “β2” in FIG. 11B. Therefore,
When the designated range average value calculation processing is executed, in the example shown in FIG. 11A, the arithmetic mean is calculated as the gradation value γ1, and
In the example shown in FIG. 11B, the arithmetic mean is calculated as the gradation value γ2.

The threshold value determination processing including the specified range average value calculation processing is executed by the following procedures (1) to (5). Steps (1) to (5) will be described in detail with reference to FIG. 4. Even when the information code area in the image data is extremely smaller than the background area, the threshold is set appropriately. Explain that Note that FIG.
The distribution is similar to that of FIG.

First, in step (1), the gradation values "0" to "255" of the entire range are designated, and the arithmetic mean calculated in the designated range average value calculation processing is used as the reference value. In this case, as shown in FIG. 4A, a lower gradation value α and an upper gradation value β are calculated, and the arithmetic mean γ is used as a reference value.

In the following procedure (2), a gradation value equal to or smaller than the reference value γ is designated, that is, gradation values “0” to “γ” are designated, and a range average value calculation process is executed. In this case, FIG.
As shown in (b), the lower gradation value a1 and the upper gradation value b
1 is calculated, and the arithmetic mean c1 is calculated. Then, the arithmetic mean c1 becomes the dark side average value.

In step (2), a tone value equal to or greater than the reference value γ is designated, that is, tone values “γ” to “255”.
Is specified, and the range average value calculation process is executed. In this case, as shown in FIG. 4B, the lower gradation value a2 and the upper gradation value b2 are calculated, and the arithmetic mean c2 is calculated. Then, the arithmetic mean c2 becomes a light-side average value.

In the following step (3), as shown in FIG. 4B, the arithmetic mean γ ′ of the dark side average value c1 and the light side average value c2 obtained in step (2) is obtained as a boundary value. . And
In step (4), the boundary value γ ′ is compared with the reference value γ obtained in (1). When the difference between the boundary value γ ′ and the reference value γ is equal to or smaller than a predetermined value, the boundary value γ ′ is determined as the threshold γ ′. On the other hand, when the difference between the boundary value γ ′ and the reference value γ is larger than a predetermined value, the boundary value γ ′ is set as a new reference value and the above (2)
Repeat steps from.

That is, when the procedure from the above (2) is repeated, the average value on the dark side approaches the average gradation value of the distribution of pixels corresponding to “dark”, while the average value on the light side is “bright”. ”Approaches the average gradation value of the pixel distribution corresponding to“. Therefore,
The boundary value calculated in step (3) gradually converges to a certain value, and the threshold is determined in step (4).

Conventionally, there has been a method in which a gradation value that includes 10% of the dark side and 10% of the bright side in all the pixels is obtained, and the arithmetic mean thereof is used as a threshold value. However, in this case, for example, when the area occupied by the information code in the image is extremely small compared to the background image, that is, when the pixels corresponding to “dark” are extremely small compared to the pixels corresponding to “bright”. As shown in FIG. 11B, the calculated threshold value γ2 is shifted to a higher gradation value (brighter side), and as a result,
The pixel corresponding to "bright" (indicated by symbol A in FIG. 11B) is inconveniently binarized as "dark".

On the other hand, in the present invention, a pixel below the boundary value (reference value) is regarded as a pixel on the "dark" side, and an average gradation value (average value on the dark side) of the pixel on the "dark" side is obtained. At the same time, a pixel which is equal to or larger than the boundary value is defined as a pixel on the “bright” side, and an average tone value (bright side average value) of the pixel on the “bright” side is calculated (the above procedure (2)). Is determined as a new boundary value, and the threshold is determined (the above procedures (3) and (4)). In other words, the pixels on the “dark” side and “bright”
Since the average tone value of each of the pixels on the side is calculated and processed, even if the number of pixels on the “dark” side and the number of pixels on the “bright” side are extremely different, ”And a pixel corresponding to“ bright ”can be determined appropriately. As a result, even if the area ratio between the information code area and the background area in the captured image data changes, the multi-value data can be appropriately binarized.

In calculating the average tone value of the pixels on the "dark" or "bright" side, the tone value including a predetermined percentage of the pixels having the lower and higher tone values among the target pixels. (Lower gradation value, upper gradation value) is calculated, and the average gradation value of the pixels on the “dark” or “bright” side is calculated. Therefore, even when the mode does not exist in the distribution of the pixels corresponding to “dark” or “bright”, the threshold can be appropriately determined.

The above-mentioned method is designed to divide the pixels of the entire image data into pixels on the "dark" side and pixels on the "bright" side. The number should be equal in an appropriate range before and after the gradation value. As can be seen from FIG.
This is because the number of pixels is flat in a range between the distribution of pixels corresponding to “dark” and the distribution of pixels corresponding to “bright”. Based on this premise, it is conceivable to adopt, for example, a binarization method as described in claim 2.

In the binarization method of the present invention, the basic procedure for binarization is the same as that of the first aspect. Since only the threshold value determining process is different, the threshold value determining process will be described. The threshold value determination processing in the binarization method of the present invention includes an average value calculation processing.
Therefore, the average value calculation processing will be described first.

In the average value calculation processing, a lower gradation value and an upper gradation value are obtained, and an arithmetic average of both the gradation values is obtained. The lower gradation value is a gradation value smaller than the designated gradation value, and is a gradation value such that a predetermined number of pixels are included in a range from the lower gradation value to the designated gradation value. Refers to the key value. On the other hand, the upper gradation value is a gradation value larger than the designated gradation value and includes a predetermined number of pixels in a range from the designated gradation value to the upper gradation value. Value.

For example, FIG. 6 shows the threshold value γ shown in FIG.
FIG. 5 is an explanatory diagram showing an enlarged correspondence relationship between the number of pixels in the vicinity and a gradation value. Note that the gradation value γ in FIG.
This corresponds to the gradation value γ in (a). In FIG.
When the gradation value γ is designated, if a predetermined number of pixels among the pixels having the designated gradation value γ or less fall within the range D, the gradation value a becomes the lower gradation value. On the other hand, if a predetermined number of pixels among the pixels having the designated gradation value γ or more fall within the range C, the gradation value b becomes the upper gradation value. Therefore, if the average value calculation process is executed with the gradation value γ specified, the arithmetic mean γ ′ of the lower gradation value a and the upper gradation value b is calculated.

The threshold value determination processing including the average value calculation processing is executed in the following procedures (1) to (3).
Hereinafter, the procedures (1) to (3) will be specifically described. First, in the procedure (1), the maximum gradation value of a predetermined proportion of pixels having a relatively small gradation value among all pixels and the predetermined proportion of a relatively large gradation value of all pixels And the minimum gradation value of the pixel is obtained, and the arithmetic mean of the gradation values is used as a reference value. This processing is the same as the procedure (1) of the binarization method according to the first aspect.

In the following procedure (2), the reference value calculated in the procedure (1) is designated, and the arithmetic mean calculated in the average value calculation processing is set as a boundary value. Specifically, procedure (1)
Then, as shown in FIG. 4A, the gradation value α and the gradation value β
, And an arithmetic mean γ is calculated and used as a reference value. In the procedure (2), as shown in FIG. 6, a lower gradation value a and an upper gradation value b including a predetermined number of pixels before and after the reference value γ are obtained by an average value calculation process. Arithmetic mean γ '
Is calculated as a boundary value.

In the procedure (3), when the difference between the boundary value and the reference value is equal to or smaller than a predetermined value,
The boundary value is determined as a threshold. On the other hand, when the difference between the boundary value and the reference value is larger than a predetermined value, the procedure from (2) is repeated using the boundary value as a new reference value. The case where the difference between the boundary value γ ′ and the reference value γ is equal to or smaller than a predetermined value refers to the case where the boundary value γ ′ has converged to some extent.

That is, if the procedures (2) and (3) are repeated, the boundary value γ ′ moves until the distribution of the number of pixels with respect to the gradation value becomes constant in an appropriate range before and after the boundary value γ ′. become. In FIG. 4, the boundary value γ ′ moves to the left (the pixel direction on the “dark” side) and converges to the reference value γ.

The present invention focuses on the fact that the number of pixels is constant in a range between the distribution of pixels corresponding to "dark" and the distribution of pixels corresponding to "bright". The threshold is corrected according to the distribution of pixels near the threshold. With this, even if the number of pixels on the “dark” side and the number of pixels on the “bright” side are extremely different, a threshold value that appropriately separates pixels corresponding to “dark” and pixels corresponding to “bright” can be obtained. Can be determined. As a result, even if the area ratio between the information code area and the background area in the captured image data changes, the multi-value data can be appropriately binarized.

Further, since the threshold value is determined in accordance with the distribution of pixels near the gradation value, the threshold value is determined appropriately even when the mode does not exist in the distribution of pixels on the "dark" or "bright" side. be able to. Note that the predetermined number in the average value calculation process may be determined according to the difference between the boundary value and the reference value. When the difference between the boundary value and the reference value is small, the predetermined number is increased. Conversely, when the difference between the boundary value and the reference value is large, the predetermined number is reduced. By increasing the predetermined number or decreasing the predetermined number,
This is because the convergence time of the calculated boundary value changes. With this configuration, the boundary value calculated in the procedure (2) can be converged at an appropriate speed.

According to the first aspect of the present invention, the pixels are divided into pixels on the "dark" side and pixels on the "bright" side. , An appropriate threshold was determined.
In other words, it is assumed that only the distribution of pixels on the “dark” side and the distribution of pixels on the “bright” side exist in the overall distribution of pixels. However, for example, a case may be considered in which a symbol is printed on the information recording surface on which the information code is printed so as to overlap the information code. It is conceivable that this symbol is printed in light blue, for example, and in that case, the pixel corresponding to the symbol portion is a floor around the middle of the pixel corresponding to “bright” and the pixel corresponding to “dark”. It is possible to have a tonal value. For example, FIG. 8B shows a symbol B
A distribution as shown by appears.

It is desired to binarize the symbol portion as "bright" in order to separate it from the information code. For example, according to the method described in claim 1, a pixel corresponding to the symbol portion is "dark". It is also conceivable that the pixel is regarded as the pixel on the "" side. In this case, the pixel corresponding to this symbol is “dark”
May be binarized. According to the method described in claim 2, for example, as can be seen from FIG. 8B, the threshold value is set at the boundary between the distribution of pixels corresponding to “bright” and the distribution of pixels corresponding to the design portion. May be determined. Therefore, also in this case, the pixel corresponding to the design is
It may be binarized as “dark”.

Therefore, it is conceivable to adopt a method as described in claim 3. Also in the present invention, the basic procedure for performing binarization is the same as in the first and second aspects of the present invention. Since the threshold value determination processing is different, the threshold value determination processing will be described.

The threshold value determination processing in the binarization method of the present invention is executed according to the following procedures (1) to (3). Procedures (1) to (3) will be specifically described with reference to FIGS. In FIG. 8A, a description is given on the assumption that a halftone pattern is printed on the entire image data.

First, in step (1), the entire image data is
The image is divided into small areas including a predetermined number of pixels, and the difference between the maximum gradation value and the minimum gradation value of the pixels included in the small area is calculated. For example, the entire image data is divided into small areas e1, e2, e3,... As shown in FIG. Note that, for example, one small area may be a 3 × 3 pixel area. And each small area e
The difference between the maximum gradation value and the minimum gradation value of the pixel included in each of the pixels 1, e2, e3,... Is calculated. In the small areas e8 and e9 shown in FIG. 8A, the difference between the maximum gradation value and the minimum gradation value is large because a part of the information code is in a part of the area. On the other hand, small areas e1 to e
7, the difference between the tone value of the pixel corresponding to the halftone pattern portion and the tone value of the "bright" pixel is smaller than that of the small areas e8 and e9. Become smaller.

In the following procedure (2), based on the distribution of the difference between the maximum gradation value and the minimum gradation value for each small area, the average of the differences is determined for the small areas distributed on the side where the difference is large.
The distribution of this difference is as shown in FIG. As described above, since the difference calculated for the small areas e1 to e7 is relatively small, these areas are distributed on the side where the gradation difference is small. On the other hand, since the difference calculated for the small areas e8 and e9 is relatively large, these areas are distributed on the side where the gradation difference is large.
The average of the differences for the small areas distributed on the side where the difference is larger corresponds to the gradation difference d2 in FIG. This gradation difference d2
Corresponds roughly to d2 in FIG. 8 (b).

In the procedure (2), the arithmetic mean of the maximum or minimum tone value of the small area included in the reference area is determined in units of a reference area composed of a predetermined number of small areas. The reference area includes a predetermined number of small areas. For example, in FIG. 8A, the reference area is composed of three small areas each of which is vertical and horizontal, and the entire image data is composed of six reference areas. The nine small areas e1 to e9 correspond to the reference area. Note that the entire area of the image data may be set as the reference area, but the threshold is determined in units of the reference area as described later, so that dividing the image data into several reference areas determines an appropriate threshold. This is advantageous in that

FIG. 8 shows the distribution of pixels for one reference area.
In the case of (b), the average of the minimum gradation values is calculated around the mode of the distribution of the pixels corresponding to “dark” (indicated by the symbol h1). On the other hand, the average of the maximum gradation values is calculated around the mode of the distribution of pixels corresponding to “bright” (indicated by the symbol h2).

In the following procedure (3), a threshold value is determined by correcting the arithmetic mean of the minimum or maximum gradation value by the average of the differences. In the above procedure (2), the difference is calculated as the average of FIG.
The difference d2 shown in (b) is calculated. Therefore, when the minimum gradation value h1 is obtained in the procedure (2), d2 is multiplied by a predetermined value and the value added to h1 is determined as the threshold. For example, in FIG. 8B, (1) is set to the gradation value h1.
/ 4) × d2 to obtain a threshold value γ ′. On the other hand, when the maximum gradation value h2 is obtained in the procedure (2), this d2
Is multiplied by a predetermined value, and a value obtained by subtracting from h2 is determined as the threshold. For example, in FIG. 8B, the threshold value γ ′ is obtained by subtracting (3/4) × d2 from the gradation value h2.

That is, in this method, the average difference between the distribution of pixels corresponding to “dark” and the distribution of pixels corresponding to “bright” is obtained, and a threshold value is determined according to the difference. As a result, as shown in FIG. 8B, the threshold value can be determined in consideration of the distribution of pixels corresponding to a design portion that is a halftone. As a result, an appropriate threshold value can be set even for an information code printed over a symbol.

Although the invention has been described as an invention of a method for binarizing an information code, the invention can also be realized as an invention of an optical information reading apparatus using this binarization method. That is, an optical information reading apparatus is configured to read optical information using the binarization method according to any one of claims 1 to 4.

Similarly, an apparatus for judging the quality of an information code using this binarization method may be configured. That is, the present invention is a code quality determination assisting device that restores an information code to be read as image data using the binarization method according to any one of claims 1 to 4. If the print quality of the printed information code is evaluated based on the image data restored by the quality judgment assisting device, the quality of the information code can be appropriately adjusted regardless of the size of the background area in the image data. You can judge.

[0047]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. [First Embodiment] FIG. 1 is a block diagram of a two-dimensional code reader 1 according to a first embodiment. The two-dimensional code reader 1 according to the first embodiment includes a read control circuit 10, a light emitting diode (illumination LED) 11, a CCD area sensor 12, an amplification circuit 13, an A / D conversion circuit 14, The configuration mainly includes an address generation circuit 15, a synchronization pulse output circuit 16, an image memory 20, a system control circuit 30, a switch group 31, a liquid crystal display 32, and a communication I / F circuit 33.

Read control circuit 10 and system control circuit 3
0 is configured as a computer system including a CPU, a ROM, a RAM, an I / O, and the like. The read control circuit 10 executes a threshold value determination process described later according to a program stored in the ROM. On the other hand, a switch group 31, a liquid crystal display 32, and a communication I / F circuit 33 are connected to the system control circuit 30, and the system control circuit 30 mainly performs input / output control.

The illumination LED 11 emits red light for illumination to a two-dimensional code which is optical information to be read. The CCD area sensor 12 has a plurality of two-dimensionally arranged CCDs, which are light receiving elements, captures the outside world, and outputs the two-dimensional image as a horizontal scanning line signal. In the first embodiment, the scanning line signal as the “analog image signal” from the CCD area sensor 12 is amplified by the amplifier circuit 13 and output to the A / D conversion circuit 14.

The amplification circuit 13 amplifies the scanning line signal output from the CCD area sensor 12 with an amplification factor corresponding to the gain control voltage input from the reading control circuit 10. The A / D conversion circuit 14 converts an analog scanning line signal into a digital signal and outputs the digital signal to the image memory 20. The data converted into a digital signal by the A / D conversion circuit 14 is multi-value data, and the multi-value data is stored in a multi-value data storage area 20 a of the image memory 20.
In addition, in this multi-valued data, each pixel has 8 bits (“0” to
"255").

The synchronizing pulse output circuit 16 outputs a synchronizing pulse sufficiently finer than the two-dimensional image data pulse output from the CCD area sensor 12. The address generation circuit 15 counts this synchronization pulse and outputs
An address for the multi-value data storage area 20a of 0 is generated. The multilevel data is written in 8-bit units for each address.

The read control circuit 10 reads out multi-value data stored in the multi-value data storage area 20a of the image memory 20, and performs a threshold value determination process for determining a threshold value for binarization based on the multi-value data. Execute. Then, based on the threshold value determined in the threshold value determination process, the read multi-value data is converted into binary data (binarization), and stored in the binary data storage area 20b of the image memory 20 again.

The switch group 31 includes a reading switch for instructing the user to start reading processing, a numeric keypad, and various function keys, and is used for inputting information. The liquid crystal display 32 is configured as, for example, a two-tone LCD, and is used for displaying read optical information and the like.

The communication I / F circuit 33 communicates with an external device (not shown), and transmits data to the external device via a communication light emitting element (not shown), or transmits light to a communication light receiving element (not shown). (For example, a program for operating the system and an instruction to wait for transmission) from the external device.

The two-dimensional code reader 1 according to the first embodiment having the above-described configuration amplifies the scanning line signal output from the CCD area sensor 12 by the amplifier circuit 13, and amplifies the amplified scanning line signal. Is converted into multi-value data by the A / D conversion circuit 14 and stored in the multi-value data storage area 20a of the image memory 20. Then, the reading control circuit 1
0 executes a threshold value determination process based on the multi-value data stored in the multi-value data storage area 20a. Then, the multi-valued data is converted to 2 by the threshold value determined in the threshold value determination process.
It is converted into a value and stored as binary data in a binary data storage area 20a of the image memory 20. Then, the reading control circuit 10
Executes the reading process of the optical information, and the image memory 20
The optical information is read based on the binary data stored in the binary data storage area 20a.

The two-dimensional code reader 1 of the first embodiment
Has a feature in the above-described threshold value determination processing. Therefore, next, a threshold value determining process executed by the reading control circuit 10 will be described with reference to the drawings. Since the threshold value determination processing includes the specified range average value calculation processing, the specified range average value calculation processing will be described first based on the flowchart of FIG. The threshold value determination processing will be described based on a flowchart.

In the first step S300 in FIG. 3, a pixel satisfying a predetermined condition is selected from pixels having a gradation value in a specified range. The specified range average value calculation process is called from a threshold value determination process described later, and a range of gradation values is specified in the threshold value determination process. Therefore, here, the processing is performed on the pixels having the gradation values in the specified range. Specifically, 10% of the pixels having relatively small gradation values are selected from the pixels having the gradation values in the specified range.

At S310, the maximum gradation value among the gradation values of the pixel selected at S300 is set as the lower gradation value. In the next step S320, a pixel satisfying a predetermined condition is selected from the pixels having the gradation value in the specified range. Here, 10% of pixels having relatively large gradation values are selected from the pixels having gradation values in the specified range.

At S330, the smallest gradation value among the gradation values of the pixel selected at S320 is set as the upper gradation value. Then, in the next S340, an arithmetic mean of the lower gradation value calculated in S310 and the upper gradation value calculated in S330 is obtained, and thereafter, the process returns to the processing step of the caller.

The above is the designated range average value calculation processing. Subsequently, the threshold value determination processing will be described based on the flowchart of FIG. First, in the first step S100, the gradation values "0" to "255" of the entire range are designated. The following S1
At 10, the designated range average value calculation processing described above with reference to FIG. 3 is called. Then, in the next S120, the arithmetic mean γ obtained in the designated range average value calculation processing is set as the reference value.

Here, the processing of S100 to S120 will be specifically described with reference to FIG. Image memory 2 as described above
The multi-valued data stored in the 0-valued multi-value data storage area 20a is data of pixels having gradation values of “0” to “255”. FIG. 4 is an explanatory diagram showing gradation values of certain multi-value data and distribution of pixels having the gradation values.

At S100, the gradation values "0" to "2"
When “55” is designated and the designated range average value calculation process is called in S110, in the designated range average value calculation process, first, the relative range among the pixels having the gradation values “0” to “255” in the designated range is used. 10% of pixels having an extremely small gradation value are selected (S300). In FIG. 4A, pixels distributed in the range of gradation values “0” to “α” correspond to relatively small pixels of 10%. Then, the maximum gradation value of these pixels is set as the lower gradation value (S310). That is, the lower gradation value is the gradation value “α”.

Subsequently, the gradation values "0" to "25" in the designated range are set.
5% of the pixels having a relatively large gradation value among the pixels having "5" are selected (S320). In FIG. 4, the pixels distributed in the range of the gradation values “β” to “255” correspond to relatively large 10% of the pixels. Then, the minimum gradation value of these pixels is set as the upper gradation value. That is, the upper gradation value is the gradation value “β”.

Then, an arithmetic mean γ of the lower gradation value α and the upper gradation value β is obtained (S340). S12 described above
At 0, this arithmetic mean γ is used as a reference value. Following S13
In the case of 0, a gradation value equal to or less than the reference value is designated, and in the next S140, the designated range average value calculation processing is called. Then, the arithmetic mean obtained in the designated range average value calculation processing is set as a dark side average value.

In the next step S160, a gradation value equal to or larger than the reference value is designated, and in the next step S170, the designated range average value calculation processing is called. Then, the arithmetic mean obtained in the designated range average value calculation processing is set as the light side average value. And the following S
At 190, the arithmetic mean of the dark side average value and the light side average value is calculated as a boundary value.

Here, the processing of S130 to S190 will be specifically described with reference to FIG. Gradation value "0" to below the reference value
When “γ” is designated (S130) and the designated range average value calculation process is executed (S140), as shown in FIG. 4B, the lower gradation value a1 and the upper gradation value b1 are calculated, and the phase is calculated. The weighted average c1 is calculated. Then, the arithmetic mean c1 becomes the dark side average value (S150). Further, when the gradation values “γ” to “255” which are equal to or larger than the reference value are designated (S160), and the range average value calculation processing is executed (S170), as shown in FIG. a2 and an upper gradation value b2 are calculated, and an arithmetic mean c2 is calculated. Then, the arithmetic mean c2 becomes a light-side average value (S180). Then, the arithmetic mean γ ′ of the dark side average value c1 and the light side average value c2 is calculated as a boundary value (S190).

In S200 in FIG. 2, it is determined whether or not the difference between the boundary value γ ′ and the reference value γ is equal to or smaller than a predetermined value ε. Here, when | γ′−γ | ≦ ε (S200: YE
S), the threshold value is set to the boundary value γ ′ in S210, and then the threshold value determination processing ends. On the other hand, if | γ′−γ |> ε (S200: NO), the process from S130 is repeated with the boundary value γ ′ as a new reference value γ in S220.

When the processing of S130 to S190 is repeated, the dark side average value approaches the average gradation value of the distribution of pixels corresponding to “dark”, while the light side average value corresponds to “bright”. It approaches the average gradation value of the pixel distribution. Therefore, the boundary value γ ′ gradually converges to a certain gradation value.

Conventionally, there has been a method in which a gradation value that includes 10% of the dark side and 10% of the bright side in all the pixels is obtained, and the arithmetic mean thereof is used as a threshold value. However, when the area occupied by the information code in the image becomes extremely small, that is, when the number of pixels on the dark side becomes extremely small as compared with the number of pixels on the light side, the calculated threshold is calculated as shown in FIG. γ2
Shifts to the higher (brighter) side of the tone value,
The pixel corresponding to "bright" (indicated by symbol A in FIG. 11B) is inconveniently binarized as "dark".

On the other hand, in the first embodiment, by specifying a gradation value equal to or smaller than the reference value (S130 in FIG. 2), the designated range average value is calculated (S14).
0), the average tone value (dark-side average value) of pixels on the “dark” side is obtained (S150), and a tone value equal to or higher than the reference value is designated (S1)
60) By executing a designated range average value calculation process (S170), an average gradation value (bright side average value) of pixels on the “bright” side is obtained (S180), and the arithmetic mean thereof is calculated as a new boundary value. The process of setting (reference value) (S220) is repeated to determine the threshold value (S210). In other words, the pixels on the “dark” side and the pixels on the “bright” side are separated, and the threshold value is calculated by calculating the average tone value of each. Even if the number of pixels is extremely different,
It is possible to determine a threshold value that appropriately separates a pixel corresponding to “dark” from a pixel corresponding to “bright”. As a result, even if the area ratio between the two-dimensional code area and the background area in the captured image changes, the multi-value data can be appropriately binarized.

In the designated range average value calculation processing shown in FIG. 3, when calculating the average gradation value of the pixels on the “dark” or “bright” side, the lower gradation value of the target pixels is used. And the gradation value (lower gradation value, upper gradation value) of the boundary including a predetermined percentage of pixels on the higher side is calculated (S310, S310 in FIG. 3).
330), an average gradation value of pixels on the “dark” or “bright” side is obtained (S340). Therefore, even when the mode does not exist in the distribution of the pixels corresponding to “dark” or “bright”, the threshold can be appropriately determined. [Second Embodiment] In the first embodiment, all the pixels are “dark”.
The threshold value is determined from the average value of the distribution of the pixels on the “dark” side and the average value of the distribution of the pixels on the “bright” side.

On the other hand, the number of pixels near the gradation value to be determined as the threshold value should be equal in an appropriate range before and after the gradation value. As can be seen from FIG. 4A, the distribution of pixels is flat in the range between the distribution of pixels corresponding to “dark” and the distribution of pixels corresponding to “bright”. An embodiment in which the threshold is determined based on this premise will be described below as a second embodiment. In addition,
The description of the same parts as those in the first embodiment will be omitted, and the same reference numerals will be given. Here, the first
Only the threshold value determination processing different from the embodiment will be described.

FIG. 5 is a flowchart showing a threshold value determining process executed by the reading control circuit 10 in the second embodiment. First, in the first step S400,
The gradation values “0” to “255” of the entire range are designated. The following S
At 410, the designated range average value calculation processing described above with reference to FIG. 3 in the first embodiment is called. by this,
As in the first embodiment, an arithmetic mean γ of the lower gradation value α and the upper gradation value β as shown in FIG. In subsequent S420, the average value γ obtained in S410 is set as a reference value.

Then, in S430, the “lower gradation value” is set so that the number of pixels having the gradation value “a” equal to or smaller than the reference value γ and in the range of gradation values “a” to “γ” is a predetermined number K. The tone value “a” as the “value” is obtained. Similarly, next in S440,
A gradation value b equal to or more than the reference value γ, and the gradation value “γ”
A gradation value “b” as an “upper gradation value” such that a predetermined number K of pixels in a range up to “b” is obtained.

In S450, the arithmetic mean of the upper gradation value a and the lower gradation value b is calculated as the boundary value γ '. In addition,
The processing of S430 to S450 described above corresponds to “average value calculation processing”. Here, the processing of S430 to S450 is
A specific description will be given with reference to FIG.

For example, FIG. 6 is an explanatory diagram in which the correspondence between the number of pixels near the tone value γ shown in FIG. 4A and the tone value is enlarged. The tone value γ in FIG. 6 corresponds to the tone value γ in FIG. In FIG. 6, the gradation value γ
> A, and a predetermined number K of pixels distributed in a range of gradation values “a” to “γ” (a range indicated by a symbol D in FIG. 6) is obtained. (S430). The predetermined number K corresponds to the lower area in the range D of the distribution curve in FIG. On the other hand, when the gradation value γ <b, the gradation value “γ”
The gradation value “b” is calculated so that the number of pixels in the range of “−b” (the range indicated by the symbol C in FIG. 6) is a predetermined number K (S4).
40). The predetermined number K corresponds to the lower area in the range C of the distribution curve in FIG. Then, the arithmetic mean of the gradation value a and the gradation value b is calculated as the boundary value γ ′ (S45).
0).

Then, in S460 in FIG. 5, it is determined whether or not the difference between the boundary value γ ′ and the reference value γ is equal to or smaller than a predetermined value ε. Here, when | γ′−γ | ≦ ε (S46
0: YES), the boundary value γ 'is set as a threshold value in S470, and the threshold value determination processing is thereafter terminated. On the other hand, | γ'-γ |
If> ε (S460: NO), the process from S430 is repeated in S480 using the boundary value γ ′ as a new reference value γ.

By repeating the processing from S430 to S450,
The boundary value γ ′ moves until the distribution of the number of pixels with respect to the gradation value becomes constant in an appropriate range before and after the boundary value γ ′. In FIG. 4, the boundary value γ ′ moves to the left (pixel direction on the “dark” side) from the reference value γ and converges.

In the second embodiment, the boundary between the distribution of pixels corresponding to “dark” and the distribution of pixels corresponding to “bright” is focused on the fact that the distribution of pixels is constant. The threshold is determined according to the distribution of the pixels near the value. With this, even if the number of pixels on the “dark” side and the number of pixels on the “bright” side are extremely different, a threshold value that appropriately separates pixels corresponding to “dark” and pixels corresponding to “bright” can be obtained. Can be determined. As a result, even if the area ratio between the two-dimensional code area and the background area in the captured image changes, the multi-value data can be appropriately binarized.

Further, since the threshold value is determined in accordance with the distribution of pixels in the vicinity of the gradation value, the threshold value is determined appropriately even when the mode value does not exist in the distribution of pixels on the “dark” or “bright” side. be able to. [Third Embodiment] In the first embodiment, all pixels are divided into "dark" pixels and "bright" pixels. In the second embodiment, pixels near the boundary value are divided. The appropriate threshold value is determined by paying attention to the distribution of. In other words, it is assumed that only the distribution of pixels on the “dark” side and the distribution of pixels on the “bright” side exist in the overall distribution of pixels. However, for example, a case may be considered in which a symbol is printed on the information recording surface on which the information code is printed so as to overlap the information code. This design may be printed in light blue, for example, in which case,
It is conceivable that the pixel corresponding to the symbol portion has a gradation value around the middle of the pixel corresponding to “bright” and the pixel corresponding to “dark”. For example, as shown by a symbol B in FIG. 8B, the distribution of the pixels corresponding to the symbol part appears between the distribution of the pixels corresponding to “dark” and the distribution of the pixels corresponding to “bright”. It is.

It is desired to binarize this symbol portion as “bright” in order to separate it from the information code. According to the method described in the first embodiment, the distribution of pixels including the symbol portion is reduced. It is also conceivable that the pixel is regarded as a “dark” side pixel. In this case, the pixel corresponding to this symbol is
It may be binarized as “dark”. According to the method described in the second embodiment, as can be seen from FIG. 8B, the threshold value is set at the boundary between the distribution of pixels corresponding to “bright” and the distribution of pixels corresponding to the design portion. May be determined. Therefore, also in this case, the pixel corresponding to the symbol may be binarized as “dark”.

The two-dimensional code reader 1 of the third embodiment
Determines the appropriate threshold value and binarizes the information code even when the symbol is printed on the recording surface of the information code so as to overlap the information code. The description of the same parts as those in the first and second embodiments will be omitted, and the description will be made using the same reference numerals. Here, only the threshold value determination processing of the first and second embodiments and the subsequent binarization processing corresponding to the binarization processing will be described.

FIG. 7 is a flowchart showing the binarization processing in the third embodiment. First step S
At 500, the captured multi-valued image data is divided into small areas. In the third embodiment, one small area is 3
× 3 pixels. In subsequent S510, the maximum gradation value and the minimum gradation value of the pixels included in each small area and the difference between them are calculated.

In the next S520, a reference area for three small areas in the vertical and horizontal directions is selected, and the arithmetic mean h2 of the maximum gradation value of the small area included in the reference area is determined using the reference area as a unit. Then, in S530, the distribution of the difference between the maximum gradation value and the minimum gradation value of each small area is obtained. In S540, the average value d2 of the distribution having a relatively large difference in the difference distribution is calculated.

At S550, a value obtained by subtracting (3/4) times d2 from the average value h2 calculated at S520 is calculated as a threshold value. Then, in S560, the pixels included in each small area of the selected reference area are binarized using the threshold calculated in S550.

At S570, it is determined whether all the areas have been binarized. Here, when there is an area that has not been binarized (S570: NO), the processing from S520 is repeated. On the other hand, when it is determined that all the areas have been binarized (S570: YES), the binarization processing ends.

Here, 2 in the third embodiment described above.
The value conversion process will be specifically described with reference to FIG. As described above, the entire image data is divided into small areas including a predetermined number of pixels (S500 in FIG. 7), and the maximum gradation value, the minimum gradation value, and the minimum gradation value of the pixels included in the small area are determined. The difference is calculated (S510). That is, as shown in FIG. 8A, the entire image data is divided into small areas e1, e2, e3,.
· Divide into And each small area e1, e2, e
The maximum gradation value and the minimum gradation value of the pixel included in each of 3,... Are obtained, and the difference is calculated. In the small areas e8 and e9 shown in FIG. 8A, the difference between the maximum gradation value and the minimum gradation value is large because a part of the information code is in a part of the area. On the other hand, small area e
In the case of 1 to e7, the difference between the tone value of the pixel corresponding to the halftone pattern portion and the tone value of the "bright" pixel is obtained.
The difference between the maximum gradation value and the minimum gradation value is smaller than in e9.

Next, the arithmetic mean of the maximum gradation value of the small area included in the reference area is determined in units of a reference area composed of a predetermined number of small areas (S520). This reference area is composed of three vertical and horizontal small areas.
As shown in FIG. 8A, the entire area is divided into six reference areas. For example, nine small areas e1 to e9 constitute a reference area. Then, the arithmetic mean of the maximum gradation value of the small area included in the reference area is obtained. The reference area may be, for example, the entire area of the image data. However, since the threshold is determined in units of the reference area as described later, it is more appropriate to divide the image data into several reference areas. This is advantageous in that a suitable threshold is determined.

FIG. 8 shows the distribution of pixels with respect to the reference area.
In the case of (b), the average of the maximum gradation value is calculated around the mode of the distribution of pixels corresponding to “bright” (indicated by the symbol h2). Next, based on the distribution of the difference calculated for each small area, an average of the differences is obtained for the small areas distributed on the side where the difference is large. The distribution of the difference is as shown in FIG. As described above, the small area e1
Since the difference calculated for .about.e7 becomes relatively small, these areas are distributed on the side where the gradation difference is small. On the other hand, since the difference calculated for the small areas e8 and e9 is relatively large, these areas are distributed on the side where the gradation difference is large. Therefore, FIG.
As can be seen from the distribution of tone value differences in the entire region, as shown in FIG.
A portion corresponding to a small area having a relatively small difference in tone value appears.

Therefore, the average gradation difference of the distribution having the larger gradation difference is calculated (S540). That is, the average value d2 in FIG. 9 is calculated. This average value d2 is shown in FIG.
It roughly corresponds to d2 in (b). Subsequently, the arithmetic mean h2 of the maximum gradation value is corrected by the average value d2 of the difference distribution to determine the threshold value. In the present embodiment, a value obtained by subtracting (3/4) times d2 from the arithmetic mean h2 is used as the threshold.

That is, the method in the third embodiment is as follows:
The difference between the distribution of pixels corresponding to “dark” and the distribution of pixels corresponding to “bright” is obtained, and a threshold is determined according to the difference.
As a result, as shown in FIG. 8B, the threshold value can be determined in consideration of the distribution of pixels corresponding to a design portion that is a halftone. As a result, an appropriate threshold value can be set even for an information code printed over a symbol.

In the third embodiment, a reference area of three small areas in the vertical and horizontal directions is selected, and the arithmetic mean of the maximum gradation values of the small areas included in the reference area is selected in units of the reference area. h2 is obtained (S520 in FIG. 7), and a value obtained by subtracting (3/4) times d2 from the arithmetic mean h2 is determined as a threshold value (S550). As shown in FIG. The arithmetic mean h1 of the minimum gradation values of the small areas included in the area may be obtained, and a value obtained by adding (１ ／) times the average d2 of the difference to the arithmetic mean h1 may be determined as the threshold. [Others] As described above, the present invention is not limited to the above embodiment at all, and can be implemented in various forms without departing from the gist of the present invention.

For example, in the first, second, and third embodiments, the present invention is applied to a two-dimensional code reader. However, it is of course possible to apply the present invention to a bar code reader for reading a bar code. It is. It is also conceivable that the present invention is applied to an apparatus for judging the quality of an information code and constitutes a code quality judgment assisting apparatus for restoring the information code to be read as image data. If the print quality of the printed information code is evaluated based on the image data restored by the code quality determination assisting device, the quality of the information code can be appropriately adjusted regardless of the size of the background area in the image data. Can be determined.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a two-dimensional code reader according to an embodiment.

FIG. 2 is a flowchart illustrating a threshold value determination process according to the first embodiment.

FIG. 3 is a flowchart illustrating a designated range average value calculation process according to the first and second embodiments.

FIG. 4 is an explanatory diagram specifically showing a threshold value determination process.

FIG. 5 is a flowchart illustrating a threshold value determination process according to the second embodiment.

FIG. 6 is an explanatory diagram specifically showing a threshold value determination process.

FIG. 7 is a flowchart illustrating a binarization process according to a third embodiment.

FIG. 8 is a flowchart illustrating a threshold value determination process in the binarization process.

FIG. 9 is an explanatory diagram showing a distribution of a gradation difference in a small area.

FIG. 10 is an explanatory diagram showing a relationship between a two-dimensional code area and a background area in an image.

FIG. 11 is an explanatory diagram showing a conventional problem.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Two-dimensional code reader 10 ... Read control circuit 11 ... Lighting diode for illumination (illumination LED) 12 ... C
CD area sensor 13 ... Amplifier circuit 14 ... A
/ D conversion circuit 15 ... address generation circuit 16 ... synchronization pulse generation circuit 20 ... image memory 20a ...
Multi-value data storage area 20b Binary data storage area 30 System control circuit 31 Switch group 32 Liquid crystal display 33 Communication I / F circuit

Claims (6)

[Claims] An image to be read is optically detected, and an analog image signal corresponding to the detected image is subjected to A / D conversion so that each pixel constituting the image has a floor corresponding to a signal level. Converts the stored multi-valued data into multi-valued data, stores the multi-valued data, performs a threshold value determining process for determining a threshold value for converting the stored multi-valued data into binary data, and calculates the threshold value determined by the threshold value determining process. In the method of binarizing an information code for converting the multi-valued data into the binary data by using the threshold value determining process, a relatively small gradation value among pixels having a gradation value in a specified range is used. The lower gradation value, which is the maximum gradation value of the predetermined ratio of pixels having a predetermined ratio, and the predetermined ratio of pixels having a relatively large gradation value among the pixels having the specified range of gradation values. Upper floor with minimum gradation value Seek and value,
An information code including a designated range average value calculation process for calculating an arithmetic mean of the lower gradation value and the upper gradation value, and executed by the following procedures (1) to (4): Binarization method. (1) A gradation value of the entire range is designated, and the arithmetic mean calculated in the designated range average value calculation processing is set as a reference value. (2) Specify a gradation value equal to or less than the reference value, specify the arithmetic mean calculated in the specified range average value calculation process as a dark side average value, and specify a gradation value equal to or more than the reference value. Then, the arithmetic mean calculated in the specified range average value calculation process is defined as a light side average value. (3) An arithmetic mean of the dark side average value and the light side average value is determined as a boundary value. (4) If the difference between the boundary value and the reference value is equal to or smaller than a predetermined value, the boundary value is determined as the threshold. On the other hand, when the difference between the boundary value and the reference value is larger than a predetermined value, the procedure from the above (2) is repeated using the boundary value as a new reference value. 2. An image to be read is optically detected, and an analog image signal corresponding to the detected image is subjected to A / D conversion so that each pixel constituting the image has a floor corresponding to a signal level. Converts the stored multi-valued data into multi-valued data, stores the multi-valued data, performs a threshold value determining process for determining a threshold value for converting the stored multi-valued data into binary data, and calculates the threshold value determined by the threshold value determining process. In the binarization method of an information code for converting the multi-valued data into the binary data by using the threshold value determination process, the threshold value determination process is performed with a gradation value smaller than a designated gradation value (lower gradation value). Thus, a lower gradation value such that a predetermined number of pixels are included in the range from the lower gradation value to the designated gradation value, and a gradation value larger than the designated gradation value (upper side). Gradation value), which is higher than the specified gradation value. An average value calculating process for obtaining an upper grayscale value such that a predetermined number of pixels are included in a range up to the grayscale value, and calculating an arithmetic average of the lower grayscale value and the upper grayscale value. An information code binarization method, which is executed by the following procedures (1) to (3). (1) A lower gradation value which is a maximum gradation value of a predetermined percentage of pixels having a relatively small gradation value among all pixels, and a relatively large gradation value among all pixels An upper gradation value which is a minimum gradation value of a predetermined percentage of pixels is obtained, and an arithmetic average of the lower gradation value and the upper gradation value is obtained as a reference value. (2) The reference value is specified, and the arithmetic mean calculated in the average value calculation processing is set as a boundary value. (3) If the difference between the boundary value and the reference value is equal to or smaller than a predetermined value, the boundary value is determined as the threshold. On the other hand, when the difference between the boundary value and the reference value is larger than a predetermined value, the procedure from the above (2) is repeated using the boundary value as a new reference value. 3. The binarization method according to claim 2, wherein the predetermined number in the average value calculation processing is determined according to a difference between the boundary value and the reference value. A binarization method for an information code, characterized in that: 4. An image to be read is optically detected, and an analog image signal corresponding to the detected image is subjected to A / D conversion so that each pixel constituting the image has a floor corresponding to a signal level. Converts the stored multi-valued data into multi-valued data, stores the multi-valued data, performs a threshold value determining process for determining a threshold value for converting the stored multi-valued data into binary data, and calculates the threshold value determined by the threshold value determining process. In the binarization method of an information code for converting the multi-valued data into the binary data using the information code, the threshold value determination processing is executed by the following procedures (1) to (3). Binarization method. (1) Divide into small areas including a predetermined number of pixels, and calculate the difference between the maximum gradation value and the minimum gradation value of the pixels included in the small area. (2) Based on the distribution of the small areas with respect to the difference of the gradation values, an average gradation difference of the distribution of the small areas having a relatively large difference is obtained, and a reference composed of a predetermined number of small areas. The arithmetic mean of the maximum or minimum gradation value of the small area included in the reference area is determined in units of the area. (3) The arithmetic mean is corrected based on the average gradation difference to determine a threshold value for each of the reference areas. 5. An optical information reading apparatus configured to read optical information by using the binarization method according to claim 1. 6. A code quality judgment assisting device for restoring an information code to be read as image data using the binarization method according to claim 1.