JP2002033912A - Image-processing method and image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image-processing method and an image processor, capable of detecting a watermark pattern which is hardly affected by illumination nonuniformity or sheet thickness nonuniformity or by any print pattern, such as a printed character or pattern. SOLUTION: The method includes the steps in which the image of a gift certificate is inputted in a step ST1, and the maximum value filter processing is applied to the inputted image in a step ST2; edge detection processing is applied to the image to which the maximum value filter processing is operated in the step ST2 in a step ST3; the mean value projection pattern to a vertical axis is generated to the image to which the edge detection processing is applied in the step ST3 in a step ST4; a power spectral is calculated for the mean value projection pattern generated in the step ST4 in a step ST5; and the presence or absence of watermark is judged, base don the power spectra which has been calculated in the step ST4 in a step ST6.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a method for determining whether a specific pattern such as the watermark pattern is present in a printed matter such as a security having a specific pattern such as a watermark pattern drawn by changing the thickness of a sheet. The present invention relates to an image processing method and an image processing device for detecting whether the image processing is performed.

[0002]

2. Description of the Related Art First, a specific pattern to be detected in the present invention will be described. In the present invention, for example, a specific pattern satisfying the following two conditions is set as a detection target.

(1) For example, as shown in FIG. 11, a rectangular pattern P1 provided on a printed matter S1 is provided.
A plurality are arranged periodically.

(2) A pixel corresponding to a specific pattern has a higher (or lower) luminance value as compared with peripheral pixels.

The present invention detects whether or not such a specific pattern exists in a printed matter.

A specific example of the specific pattern satisfying the above two conditions is a watermark pattern. The watermark pattern is a pattern drawn on a securities such as a bill, a check, a gift certificate, or the like by a change in the thickness of a sheet to prevent forgery or the like. Here, a specific example of an image including a watermark pattern is shown in FIG.

FIG. 12 shows, for example, an image of a gift certificate S2, which has a printed pattern P2 such as a printed character or pattern, and which has a higher brightness value than the background in the printed pattern P2. It has a plurality of rectangular watermark patterns P3 that are periodically arranged. Although illustration is omitted, it is assumed that a background pattern is also printed. Hereinafter, detection of such a watermark pattern will be described as a specific example.

The following two typical methods are known as conventional watermark pattern detection methods.

First conventional method: method based on binarization processing Second conventional method: method based on edge detection processing First, the first conventional method will be described. The first conventional method is
It consists of three steps: binarization, calculation of projection onto the vertical axis, and verification of periodicity. In the binarization of step 1, PT
Using the ile method, a threshold for binarization is determined. More specifically, a luminance histogram is calculated for the input gray image, and as shown in FIG. 13, a luminance value TH that gives an area of TAREA% from the brighter one of the distribution of luminance values of the pixels is obtained as a threshold for binarization. For an input gray image, a threshold T
“1” is assigned to a pixel having a value larger than H, and “0” is assigned to other pixels
To obtain a binary image. Thereby, the binary image S3 of the watermark pattern as shown in FIG.
Is obtained.

In the calculation of the projection onto the vertical axis in step 2,
As shown in FIG. 15, the binary image S obtained in step 1
For No. 3, a projection pattern is obtained on the vertical axis. y = Y0
Is the number of pixels having a value of "1" among the pixels of the binary image satisfying y = Y0.
The reason why the binary image S3 is processed as a two-dimensional pattern and the watermark pattern is not detected is to speed up the processing. High-speed processing can be performed by projecting the binary image on the vertical axis and converting it into a projection pattern with a small amount of data.

In the verification of the periodicity in step 3, it is determined whether or not the projection pattern obtained in step 2 has periodicity as a watermark pattern. As shown in FIG. 16, the projection pattern obtained in step 2 is divided into a range of a fixed size. The size of the range is the size of the period in the vertical axis direction of the watermark pattern known in advance. The maximum value and the minimum value are obtained for each range, and when the difference between the obtained maximum value and the minimum value is equal to or greater than a certain value, it is determined that there is a watermark pattern.

Next, a second conventional method will be described. First
The conventional method has a problem that it is susceptible to uneven lighting and uneven thickness of printed matter. In the first conventional method, fluctuations in pixel values of a watermark pattern portion in the same image become large due to uneven illumination and uneven thickness of a printed material, and the watermark pattern appears only partially in a binary image. In some cases, pixels other than the watermark pattern may appear in the binary image. In such a case, there is a possibility that the periodicity is not detected in the projection pattern onto the vertical axis, and the watermark pattern cannot be detected.

Therefore, in the second conventional method, in order to solve the problems of the first conventional method, before performing the first conventional method,
The edge detection process is performed. The second conventional method is:
Since the edge detection processing detects a pixel having a strong edge as a watermark pattern, the pixel is less affected by a change in the pixel value of the watermark pattern.

[0014]

First, problems of the first conventional method will be described. The first conventional method has a problem that it is susceptible to uneven lighting and uneven thickness of printed matter. In the first conventional method, the fluctuation of the pixel value of the watermark pattern in the same image becomes large due to uneven illumination or uneven thickness of the printed matter, and the watermark pattern appears only partially in the binary image. In some cases, pixels other than the watermark pattern may appear in the binary image. In such a case, the periodicity may not be detected in the projection pattern on the vertical axis, and the watermark pattern may not be detected.

Next, problems of the second conventional method will be described. The second conventional method is a method of first performing an edge detection process in order to solve the problem that the first conventional method is susceptible to uneven illumination and uneven thickness of a printed material. The second conventional method has a problem that a print pattern such as a character or a pattern printed on a printed matter is easily affected. Print patterns such as characters and patterns are often printed so that humans can easily see them, and the contrast between the character pattern and the paper and the contrast between the pattern and the paper are compared to the contrast between the watermark pattern and the paper. It may be large. For this reason, in an image after edge detection, a character pattern, a pattern, or the like has a large value as an edge. Therefore, even in a binary image, these patterns appear, which may adversely affect the detection of a watermark pattern. is there.

Next, problems common to the first conventional method and the second conventional method will be described. Both the first conventional method and the second conventional method have a problem that they are easily affected by high frequency components in the verification of periodicity. For example, as shown in FIG. 17, when the projection pattern obtained from the binary image includes a frequency component higher than the frequency of the watermark pattern, there is a watermark pattern regardless of whether or not there is a frequency component of the watermark pattern. Will be determined.

Accordingly, the present invention provides an image processing method and an image processing method capable of detecting unevenness in illumination and thickness of paper and a watermark pattern which is hardly affected by a printed pattern such as a printed character or pattern. It is intended to provide a device.

Another object of the present invention is to provide an image processing method and an image processing apparatus which are hardly affected by high frequency components of an input image and which can prevent erroneous determination.

[0019]

An image processing method according to the present invention has a printed pattern such as a printed character or pattern,
And an image processing method for detecting whether or not the printed pattern has a specific pattern having a specific pattern having a higher luminance value than the background in the print pattern, wherein the step of inputting an image of the printed material, Subjecting the input image to a maximum value filtering process that replaces pixels with the maximum value of neighboring pixels; and performing a maximum value filtering process on the image subjected to the maximum value filtering with a specific pattern and pixels other than the specific pattern. A step of performing edge detection processing for enhancing a boundary; and a step of determining the presence or absence of the specific pattern based on the image on which the edge detection processing has been performed.

Further, according to the image processing method of the present invention, a plurality of specific patterns having a printed pattern such as a printed character or pattern and having a higher luminance value than the background are periodically included in the printed pattern. An image processing method for detecting whether or not the specific pattern is present for the arranged printed matter,
Inputting the image of the printed matter, performing a maximum value filtering process for replacing the pixel with the maximum value of a neighboring pixel on the input image, and identifying the image that has been subjected to the maximum value filtering process. Performing an edge detection process for enhancing a boundary between the pattern and a pixel other than the specific pattern; generating an average projection pattern on the vertical axis for the image on which the edge detection process has been performed; Calculating a power spectrum for the average projection pattern; and determining whether the specific pattern is present by determining whether or not the calculated power spectrum has a local maximum value around a period of the specific pattern that is known in advance. Is provided.

Further, the image processing method of the present invention can be applied to a printed material having a printed pattern such as a printed character or pattern and having a specific pattern having a lower luminance value than the background in the printed pattern. An image processing method for detecting whether or not there is a specific pattern, wherein a step of inputting an image of the printed matter and performing a minimum value filter process on the input image to replace pixels with a minimum value of neighboring pixels. A step of performing an edge detection process on the image subjected to the minimum value filter process to emphasize a boundary between a specific pattern and a pixel other than the specific pattern, and based on the image on which the edge detection process is performed. Determining the presence or absence of a specific pattern.

Further, according to the image processing method of the present invention, a plurality of specific patterns having a printed pattern such as a printed character or pattern and having a luminance value lower than that of the background are periodically included in the printed pattern. An image processing method for detecting whether or not the specific pattern is present for the arranged printed matter,
Inputting the image of the printed matter, performing a minimum value filter process on the input image for replacing pixels with the minimum value of neighboring pixels, and identifying the image subjected to the minimum value filter process. Performing an edge detection process for enhancing a boundary between the pattern and a pixel other than the specific pattern; generating an average projection pattern on the vertical axis for the image on which the edge detection process has been performed; Calculating a power spectrum for the average projection pattern; and determining whether the specific pattern is present by determining whether or not the calculated power spectrum has a local maximum value around a period of the specific pattern that is known in advance. Is provided.

Also, the image processing apparatus of the present invention can be used for a printed matter having a printed pattern such as a printed character or pattern and having a specific pattern having a higher luminance value than the background in the printed pattern. What is claimed is: 1. An image processing apparatus for detecting whether or not a specific pattern is present, comprising: an image input unit for inputting an image of the printed matter; and replacing pixels in the image input by the image input unit with a maximum value of neighboring pixels. Maximum value filter processing means for performing maximum value filter processing, and edge detection processing for emphasizing a boundary between a specific pattern and a pixel other than the specific pattern in an image subjected to maximum value filter processing by the maximum value filter processing means Detecting means for determining whether or not the specific pattern exists based on an image on which edge detection processing has been performed by the edge detecting means. It is and means.

Further, the image processing apparatus of the present invention has a printing pattern such as a printed character or pattern, and a plurality of specific patterns having a higher brightness value than the background are periodically included in the printing pattern. An image processing apparatus for detecting whether or not the specific pattern is present for the placed printed matter,
Image input means for inputting the image of the printed matter, maximum value filter processing means for performing maximum value filter processing for replacing a pixel with a maximum value of a neighboring pixel on the image input by the image input means, and maximum value filter Edge detection means for performing edge detection processing for enhancing the boundary between a specific pattern and pixels other than the specific pattern on the image subjected to the maximum value filter processing by the processing means; and performing edge detection processing using the edge detection means. Average value projection pattern generation means for generating an average value projection pattern on the vertical axis for the image, and power spectrum calculation means for calculating a power spectrum for the average value projection pattern generated by the average value projection pattern generation means, In the power spectrum calculated by the power spectrum calculation means, It is provided with a determination means for determining whether the specific pattern around the period of the specific pattern dimethyl has been found by whether has a maximum value.

Further, the image processing apparatus according to the present invention is capable of printing a printed matter having a printed pattern such as a printed character or pattern and having a specific pattern having a lower luminance value than the background in the printed pattern. An image processing apparatus for detecting whether there is a specific pattern, comprising: an image input unit for inputting an image of the printed matter; and replacing an image input by the image input unit with a pixel having a minimum value of a neighboring pixel. Minimum value filter processing means for performing minimum value filter processing; and edge detection processing for emphasizing a boundary between a specific pattern and a pixel other than the specific pattern in an image subjected to minimum value filter processing by the minimum value filter processing means. Detecting means for determining whether or not the specific pattern exists based on an image on which edge detection processing has been performed by the edge detecting means. It is and means.

Further, the image processing apparatus of the present invention has a printing pattern such as a printed character or pattern, and a plurality of specific patterns having a lower luminance value than the background in the printing pattern. An image processing apparatus for detecting whether or not the specific pattern is present on a placed printed matter, comprising: an image input unit that inputs an image of the printed matter; and a pixel near the image input by the image input unit. Minimum value filter processing means for performing minimum value filter processing to be replaced with the minimum value of pixels, and a specific pattern and a pixel other than the specific pattern for an image subjected to minimum value filter processing by the minimum value filter processing means. Edge detection means for performing edge detection processing for emphasizing the boundary, and an average value on the vertical axis of the image on which edge detection processing has been performed by the edge detection means Mean value projection pattern generation means for generating a shadow pattern, power spectrum calculation means for calculating a power spectrum for the average value projection pattern generated by the mean value projection pattern generation means, and power calculated by the power spectrum calculation means A determination means is provided for determining the presence or absence of the specific pattern based on whether or not the spectrum has a local maximum value around the period of the specific pattern that is known in advance.

[0027]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 schematically shows a configuration of an image processing apparatus according to the present embodiment. 1, a CCD camera 11 as an image input unit captures an image of a gift certificate (printed matter) S2 as shown in FIG. 12, and inputs the image as digital image data. The image memory 12 as an image storage means stores the gift certificate S2 input by the camera 11
Is temporarily stored. The image processing unit 13 as various processing means is mainly constituted by, for example, a CPU (Central Processing Unit), and has a maximum value filter process, an edge detection process, and an average value projection pattern generation process, which will be described in detail later. , Power spectrum calculation processing, determination processing, and the like.

Next, image processing according to the embodiment of the present invention will be described with reference to the flowcharts shown in FIGS.

First, in step ST1, an image of a gift certificate S2 as shown in FIG. 12 is input by the CCD camera 11 and stored in the image memory 12.

Next, at step ST2, the image memory 1
2 is subjected to the maximum value filter processing. As shown in FIG. 4, the maximum value filtering process is a process of replacing pixels of an input image with the maximum value of neighboring pixels, and is performed for the purpose of emphasizing a watermark pattern and removing a printing pattern such as a character or a pattern. The result of performing the maximum value filter processing on the image of FIG. 12 is shown in FIG. 5, in which reference numeral P4 is a watermark pattern portion.

Next, in step ST3, step ST
In step 2, edge detection processing is performed on the image that has been subjected to the maximum value filter processing. The edge detection processing is performed to emphasize the boundary between the watermark pattern and pixels other than the watermark pattern. The result of performing the edge detection processing on the image of FIG. 5 is shown in FIG. 6, in which a symbol P5 is a watermark pattern portion.

Next, in step ST4, step ST
In step 3, an average projection pattern on the vertical axis is generated for the image subjected to the edge detection processing. The generation of the average projection pattern is, as shown in FIG. 7, a value obtained by calculating the average of the luminance values of the pixels for each line and projecting the average on the vertical axis. The average projection pattern obtained from the image of FIG. 7 is as shown in FIG.

Next, at step ST5, step ST
The power spectrum is calculated for the average projection pattern generated in step 4. The calculation of the power spectrum is performed to verify the periodicity of the waveform obtained by the projection. Here, the average projection pattern f (y) (y = 0,..., M−
Let the power spectrum of 1) be P (T) (T <0).
T represents a period. Hereinafter, the power spectrum P (T) used in the present invention is defined.

The power spectrum P (T) is the norm of the inner product of the waveform W (T, y) having a period T and f (y), and is defined as in the following equation (1).

[0036]

(Equation 1)

The waveform W (T, y) of the cycle T is a function represented by the exponential function E (y, T) of the cycle T as follows.

W (T, y) = A × E (y, T) + B Here, A and B are constants. The constants A and B are determined so that the waveform W (T, y) of the cycle T satisfies the following Expression 2 (condition 1) and Expression 3 (condition 2).

[0039]

(Equation 2)

[0040]

(Equation 3)

The above conditions 1 and 2 correspond to normalization and canonicalization of the waveform W (T, y) having the period T, respectively.

The exponential function E (y, T) of the period T is
It is assumed to be expressed as in the following Expression 4.

[0043]

(Equation 4) Here, j is an imaginary unit.

FIG. 9 shows a power spectrum obtained from the average projection pattern shown in FIG. The power spectrum of FIG. 9 has a local maximum around the period of the watermark pattern.

The power spectrum P (T) used in the present invention
Is an extension of the discrete Fourier power spectrum PDFT (T). The discrete Fourier power spectrum is defined as T = M / (M−1), M / (M−2),. On the other hand, the power spectrum P (T)
Is defined for any real number T that satisfies 0 ≦ T ≦ M, so that more detailed frequency analysis is possible.

Next, at step ST6, step ST
The presence or absence of a watermark pattern is determined based on the power spectrum P (T) calculated in step 5. In this determination process, FIG.
0, the calculated power spectrum P (T)
Then, the maximum value is determined for the range of the period of the watermark pattern that is known in advance. Here, the cycle giving the maximum value is Tmax, and the standard cycle T of the watermark pattern is Tmax.
The difference from std is obtained and compared with a predetermined threshold Thdiff. As a result, when | Tmax−Tstd | <Thdiff, it is determined that there is a watermark pattern. Otherwise, it is determined that there is no watermark pattern.

As described above, according to the above embodiment, the input image is subjected to the maximum value filter processing for replacing pixels with the maximum value of neighboring pixels, and the image subjected to the maximum value filter processing is subjected to the maximum value filter processing. The edge detection processing is performed to emphasize the boundary between the watermark pattern and the pixels other than the watermark pattern, and the presence or absence of the watermark pattern is determined based on the image on which the edge detection processing is performed. This makes it possible to detect a watermark pattern that is not easily affected by unevenness in thickness, and to detect a watermark pattern that is not easily affected by character patterns and patterns.

Further, when a plurality of watermark patterns are provided periodically, the power spectrum is used to determine the periodicity of the watermark pattern, thereby detecting the watermark pattern which is not easily affected by the high frequency components of the input image. Becomes possible.

In the above-described embodiment, the case has been described where the presence or absence of a plurality of rectangular watermark patterns having a higher luminance value than the background and arranged periodically is described.
However, the present invention is not limited to this. For example, a single watermark pattern may be used. In this case, it is unnecessary to determine the periodicity. Also, the shape of the watermark pattern is not limited to a rectangle, for example,
It may be circular or elliptical.

Furthermore, in the above-described embodiment, a case has been described in which the presence or absence of a specific pattern such as a watermark pattern having a luminance value higher than the background is detected. However, the present invention is not limited to this. The same applies to the case of detecting the presence or absence of a specific pattern having a lower luminance value than the background. In that case, the maximum value filter processing in step ST2 in FIG. 2 may be a minimum value filter processing in which pixels of the input image are replaced with minimum values of neighboring pixels, and the other configurations are the same as those in the above-described embodiment. .

[0051]

As described above in detail, according to the present invention, it is possible to detect uneven lighting and uneven thickness of paper, and a watermark pattern which is hardly affected by a printed pattern such as a printed character or pattern. Image processing method and image processing apparatus can be provided.

Further, according to the present invention, it is possible to provide an image processing method and an image processing apparatus which are hardly affected by high frequency components of an input image and which can prevent erroneous determination.

[Brief description of the drawings]

FIG. 1 is a block diagram schematically showing a configuration of an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a flow of image processing according to the embodiment of the present invention.

FIG. 3 is a flowchart showing the flowchart of FIG. 2 in more detail;

FIG. 4 is a diagram for explaining a maximum value filtering process.

FIG. 5 is a diagram showing an example of an image after a maximum value filter process has been performed.

FIG. 6 is a diagram showing an example of an image after an edge detection process has been performed.

FIG. 7 is a diagram for explaining a process of generating an average projection pattern.

FIG. 8 is a diagram illustrating an example of a generated average projection pattern.

FIG. 9 is a diagram showing an example of a calculated power spectrum.

FIG. 10 is a view for explaining determination of the periodicity of a watermark pattern from a power spectrum.

FIG. 11 is a diagram showing an example of a specific pattern to be detected in the present invention.

FIG. 12 is a diagram showing a gift certificate as a specific example of an image including a watermark pattern.

FIG. 13 is a diagram for explaining the determination of a binarization threshold value in a conventional method.

FIG. 14 is a diagram showing an example of a binary image of a watermark pattern in a conventional method.

FIG. 15 is a diagram for explaining generation of a projection pattern onto a vertical axis in a conventional method.

FIG. 16 is a diagram for explaining the determination of periodicity for a generated projection pattern in a conventional method.

FIG. 17 is a diagram illustrating an example in which a conventional method erroneously determines that there is a watermark pattern.

[Description of Signs] S2: Gift certificate (printed matter) P2: Printed pattern P3: Watermark pattern (specific pattern) 11: CCD camera (image input means) 12: Image memory (image storage means) 13: … Image processing unit (maximum value filter processing means, edge detection means, average projection pattern generation means, power spectrum calculation means, judgment means)

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06T 7/60 300 H04N 1/387 H04N 1/387 7/18 B 7/18 S 1/40 Z F term (Reference) 5B057 AA11 BA02 CA08 CA12 CA16 CE06 CE08 CE12 DA12 DB02 DB09 DC16 DC19 DC36 5C054 CA04 ED08 EJ04 FC04 FC14 HA05 5C076 AA14 AA31 BA06 5C077 LL14 MM03 PP03 PP23 PP43 PP46 PP49 PP55 PP65 PQ12 FA0318A GA55 JA11

Claims (10)

[Claims] 1. A printed matter having a printed pattern such as a printed character or pattern, and having a specific pattern having a higher luminance value than the background in the printed pattern, determines whether the specific pattern is present. An image processing method for detecting, comprising: inputting an image of the printed matter; performing a maximum value filtering process for replacing the input image with a maximum value of a neighboring pixel; and a maximum value filtering process. Performing an edge detection process for emphasizing a boundary between a specific pattern and a pixel other than the specific pattern on the image subjected to the above, and determining the presence or absence of the specific pattern based on the image subjected to the edge detection process An image processing method, comprising: 2. The method according to claim 1, wherein the print includes a print pattern such as a printed character or pattern, and a plurality of specific patterns having a luminance value higher than a background in the print pattern are periodically arranged. An image processing method for detecting whether or not there is a pattern, comprising: a step of inputting an image of the printed matter; and a step of performing a maximum value filter process on the input image for replacing a pixel with a maximum value of a neighboring pixel. Performing an edge detection process on the image subjected to the maximum value filter process to emphasize a boundary between a specific pattern and a pixel other than the specific pattern; and a vertical axis for the image subjected to the edge detection process. Generating a power spectrum pattern for the generated power spectrum pattern; calculating a power spectrum for the generated power spectrum projection pattern; Determining the presence or absence of the specific pattern based on whether or not the calculated power spectrum has a local maximum value around the period of the specific pattern that is known in advance. 3. The image processing method according to claim 1, wherein the specific pattern is a watermark pattern drawn by a change in the thickness of a printed sheet. 4. A printed matter having a printed pattern such as a printed character or pattern and having a specific pattern having a lower luminance value than the background in the printed pattern is checked for the presence or absence of the specific pattern. An image processing method for detecting, comprising: inputting an image of the printed matter; performing minimum value filtering on the input image by replacing pixels with minimum values of neighboring pixels; Performing an edge detection process for emphasizing a boundary between a specific pattern and a pixel other than the specific pattern on the image subjected to the above, and determining the presence or absence of the specific pattern based on the image subjected to the edge detection process An image processing method, comprising: 5. A printed material having a printed pattern such as a printed character or pattern, and a plurality of specific patterns having a lower luminance value than the background in the printed pattern are periodically arranged. An image processing method for detecting whether or not there is a pattern, comprising the steps of: inputting an image of the printed matter; and performing a minimum value filter process on the input image for replacing pixels with minimum values of neighboring pixels. Performing an edge detection process on the image subjected to the minimum value filter process to emphasize a boundary between a specific pattern and a pixel other than the specific pattern; and a vertical axis for the image subjected to the edge detection process. Generating a power spectrum pattern for the generated power spectrum pattern; calculating a power spectrum for the generated power spectrum projection pattern; Determining the presence or absence of the specific pattern based on whether or not the calculated power spectrum has a local maximum value around the period of the specific pattern that is known in advance. 6. A printed matter having a printed pattern such as a printed character or pattern and having a specific pattern having a higher luminance value than the background in the printed pattern is determined whether or not the specific pattern is present. An image processing apparatus for detecting, comprising: image input means for inputting an image of the printed matter; and a maximum value to be subjected to a maximum value filter process for replacing an image input by the image input means with a maximum value of neighboring pixels. Filter processing means; edge detection means for performing edge detection processing for enhancing the boundary between a specific pattern and a pixel other than the specific pattern on the image subjected to the maximum value filter processing by the maximum value filter processing means; Determining means for determining the presence or absence of the specific pattern based on the image on which the edge detection processing has been performed by the edge detection means. The image processing apparatus according to symptoms. 7. A printed material having a printed pattern such as a printed character or pattern, and a plurality of specific patterns having a higher luminance value than a background in the printed pattern are periodically arranged. An image processing apparatus for detecting whether or not there is a pattern, comprising: an image input unit for inputting an image of the printed matter; and a maximum for replacing a pixel with a maximum value of a neighboring pixel in the image input by the image input unit. Maximum value filter processing means for performing value filter processing; and edge detection processing for emphasizing a boundary between a specific pattern and a pixel other than the specific pattern in the image subjected to the maximum value filter processing by the maximum value filter processing means. Edge detecting means for performing, and an average value for generating an average projection pattern on a vertical axis for an image on which edge detection processing has been performed by the edge detecting means. Shadow pattern generating means; power spectrum calculating means for calculating a power spectrum for the average projected pattern generated by the average projected pattern generating means; and a power spectrum calculated by the power spectrum calculating means. Determining means for determining the presence or absence of the specific pattern based on whether or not the specific pattern has a local maximum value around the period of the specific pattern. 8. The image processing apparatus according to claim 6, wherein the specific pattern is a watermark pattern drawn by a change in thickness of a printed paper. 9. A printed matter having a printed pattern such as a printed character or pattern and having a specific pattern having a lower luminance value than the background in the printed pattern is checked for the presence or absence of the specific pattern. An image processing apparatus for detecting, comprising: an image input means for inputting an image of the printed matter; and a minimum value for performing minimum value filter processing for replacing a pixel with a minimum value of a neighboring pixel on the image input by the image input means. Filter processing means; edge detection means for performing edge detection processing on the image subjected to the minimum value filter processing by the minimum value filter processing means to emphasize a boundary between a specific pattern and a pixel other than the specific pattern; Determining means for determining the presence or absence of the specific pattern based on the image on which the edge detection processing has been performed by the edge detection means. The image processing apparatus according to symptoms. 10. A printed material having a printed pattern such as a printed character or pattern, and a plurality of specific patterns having a lower luminance value than the background in the printed pattern are periodically arranged. An image processing apparatus for detecting whether or not there is a pattern, comprising: an image input unit that inputs an image of the printed matter; and a minimum unit that replaces a pixel with a minimum value of a neighboring pixel in the image input by the image input unit. Minimum value filter processing means for performing value filter processing; and edge detection processing for emphasizing a boundary between a specific pattern and a pixel other than the specific pattern in the image subjected to the minimum value filter processing by the minimum value filter processing means. Edge detecting means for performing, and an average for generating an average projection pattern on a vertical axis for an image on which edge detection processing has been performed by the edge detecting means. Projection pattern generation means; power spectrum calculation means for calculating a power spectrum for the average projection pattern generated by the average projection pattern generation means; and a power spectrum calculated by the power spectrum calculation means. Determining means for determining the presence or absence of the specific pattern based on whether or not the specific pattern has a local maximum value around the period of the specific pattern.